OFFICE OF THE DEMOCRATIC LEADER Interview of CHRISTOPHER WYLIE Conducted by NANCY PELOSI and ADAM SCHIFF Wednesday, April 25, 2018 2:48 - 7:37 p.m. AUDIO TRANSCRIPTION Diversified Reporting Services, Inc. (202) 467-9200 Page 2 1 2 P R O C E E D I N G MS. PELOSI: I am going to call to order this 3 evening the interview of Christopher Wylie, former 4 director of research for Cambridge Analytica. 5 6 7 I would now like to yield to the distinguished ranking member of the committee, Mr. Schiff. MR. SCHIFF: Thank you, Madam Leader. We 8 appreciate your conducting this hearing under your 9 auspices here in your office. 10 Good afternoon, Mr. Wylie. 11 This is a transcribed interview of Christopher 12 Wylie, a former employee of Cambridge Analytica, who 13 has recently come to light as a whistleblower regarding 14 the company's alleged improper acquisition and use of 15 more than 50 million Facebook profiles. 16 17 Mr. Wylie, we are very grateful for your coming this distance to speak with us today. 18 MR. WYLIE: 19 MR. SCHIFF: It's my pleasure. As part of the ongoing criminal 20 -- congressional investigation -- criminal, too, 21 perhaps -- into Russia's interference in the 2016 U.S. 22 election, House Democratic Leader Nancy Pelosi, who was Page 3 1 also an ex officio member of the House Permanent Select 2 Committee on Intelligence, has graciously offered to 3 host us for your interview. 4 support in exploring the important allegations that 5 stem from your whistleblowing and reporting by UK's 6 Channel 4 News. 7 We greatly appreciate her We have invited our colleagues from the 8 Majority to participate today, and hope that they will 9 join us, although as yet we think that is unlikely. 10 Before we begin, I want to say a few things 11 for the record. 12 Members and staff. 13 interview, Members and staff may ask questions during 14 their allotted time period. 15 basic, but that is because we need to clarify and 16 establish certain facts and understand the full 17 dimensions of the situation. 18 we know any facts you have not previously disclosed as 19 part of any other investigation or review. 20 Questioning today will be conducted by During the course of this Some questions may seem Please don't assume that This interview will be conducted at an 21 unclassified level. We ask that you give complete 22 replies to questions based on your best recollection. Page 4 1 If a question is unclear, or you are uncertain in your 2 response, please let us know. 3 the answer to a question, or cannot remember, simply 4 say so. 5 And if you do not know You are entitled to have counsel present 6 during your interview. 7 a break, just let us know. 8 accommodation by traveling here to Washington, D.C. for 9 the interview. 10 11 If at any time you need to take We appreciate your At this time, if counsel could please state your name for the record. 12 MS. ALLEN: 13 MS. PELOSI: My name is Tamsin Allen. Ranking Member, if I just may 14 thank counsel and Mr. Wylie for being here today, thank 15 you. 16 will be able to come back. 17 And I am going to yield because I -- hopefully I MS. ALLEN: May I make a small preliminary 18 observation, which is I understand that it is -- that 19 today is being transcribed. 20 some answers that Chris would like to give, but he's 21 not able to give in public because there is an ongoing 22 criminal investigation in the UK, and we don't want to There may be something -- Page 5 1 say anything that could prejudice that investigation in 2 public. 3 So it may be that we are going to need to ask 4 to -- for those things to be excised from the 5 transcript in future. 6 that heads up. So I just wanted to give you 7 MS. PELOSI: Thank you. 8 MR. SCHIFF: Okay, thank you. 9 And this will be a closed interview, so down the road, should we 10 decide to release the transcript, we will reach out to 11 you first, and see if there are any particular 12 sensitivities that we can try to work our way through. 13 MS. ALLEN: 14 MR. SCHIFF: Thank you. All right. Well, I'm going to 15 start out with a few very basic introductory questions, 16 and then we will pass it on to my colleagues. 17 MR. WYLIE: 18 MR. SCHIFF: 19 22 If I could start at the very beginning, how did you come to work for SCL Group? 20 21 Sure. Who was a part of SCL at the time that you joined? And what can you tell us about the formation Page 6 1 of SCL? 2 MR. WYLIE: Sure. So I joined SCL Group as 3 director of research in June 2015. 4 was introduced to Alexander Nix and several other 5 senior staff members at SCL through a mutual contact 6 that I had when I previously was working for the 7 coalition government in the United Kingdom for the 8 liberal democrat side. 9 That was after I I got hired in part because the previous 10 research director died in Kenya and they had an 11 opening. 12 capacity in data analytics, cyber. 13 And also, the firm was looking to expand its It's important to understand that, firstly, 14 this was prior to Cambridge Analytica being formed. 15 the time the firm was predominantly a military 16 contractor, so its clients included the UK Ministry of 17 Defense and also the United States Department of 18 Defense, as well as several other NATO countries' 19 militaries or defense ministries. 20 At At the time several countries -- so in the 21 United States DARPA, which is the U.S. military 22 research agency, and in the United Kingdom DSTL, which Page 7 1 is DARPA's equivalent, were looking at how to better 2 understand the proliferation of ideas online and, more 3 generally, studying influence in social networks. 4 So it was quite appealing for the firm to 5 bring on somebody -- to bring on a research director 6 who had much more experience in data science, 7 analytics, and generally building profiles of people 8 based on different data sets and predicting behavior. 9 Does that answer your question, or -- 10 11 MR. SCHIFF: It does. Now, when you joined SCL, how many employees did SCL have? 12 How -- was it only in the UK? 13 Did it have a Canadian presence also, or -- 14 MR. WYLIE: Yeah. So when I joined, it was -- 15 most staff were based in London. 16 understand this is a British company, it's a British- 17 registered company. 18 joined, a very large company. 19 about, in total, amongst the entire group of companies 20 -- because there is different divisions -- about 20 or 21 so staff. 22 So it's important to There was a -- it wasn't, when I There was probably The company had a network of contacts around Page 8 1 the world. 2 had offices around the world, or had consultants or 3 contractors that were based in different countries, 4 mostly for contract acquisition. 5 6 They had partnerships with local firms, or The Canadian office that I think you are referring to was set up after I joined. 7 MR. SCHIFF: So at the time you joined, was 8 there only one office in the UK, but there were 9 employees -- 10 MR. WYLIE: 11 MR. SCHIFF: 12 There were --- in other parts of the world who were -- 13 MR. WYLIE: Yeah, it was -- there was one 14 office in London that was there, or -- let me correct 15 myself. 16 There were several offices in London for each 17 of the different divisions. In terms of offices around 18 the world, they would have contractors or people that 19 they would have on retainer who either would have a 20 small office in that country if there was ongoing 21 projects, or there would be a partner firm where one 22 person would be on retainer at a local firm with Page 9 1 contacts in that country. 2 MR. SCHIFF: And when you started at SCL, how 3 well capitalized was SCL? 4 from? 5 Where did the investors come What kind of revenues did it have at that time? MR. WYLIE: I'm not privy to the exact figures 6 at that time, so I can't give you a specific answer. 7 But it wasn't, you know, a very -- it wasn't a very 8 large -- it wasn't a large company. 9 you an exact figure of revenue. 10 MR. SCHIFF: 11 And the Mercers, did they come along after you had joined? 12 13 But I can't give MR. WYLIE: Yes. So if it's helpful, I can give you sort of a rough timeline. 14 MR. SCHIFF: 15 MR. WYLIE: Sure, that would be great. So I joined in June 2013. The -- 16 and as I said before, the sort of purpose of me joining 17 SCL Group was, essentially, to do research and 18 development, to gain better capacity in data analytics 19 and targeting and profiling online, because that's 20 something that they knew that their clientele, 21 particularly on the military side, would be interested 22 in. Page 10 1 Several months after I joined -- so this would 2 be around either the end of September or the beginning 3 of October 2013 -- is -- I believe October -- is when I 4 first met Steve Bannon. 5 and I can't give you an exact date as to when some of 6 these meetings happened, but Alexander Nix went to the 7 United States because he got introduced to Steve 8 Bannon. 9 So at the -- when I joined -- And Steve was, at the time, sort of setting up 10 Breitbart UK, so he was going back and forth between 11 the United States and the UK. 12 timing for him. 13 his ability to build what we would now call the alt- 14 right -- 15 16 17 18 19 20 21 22 So it was fortuitous He was looking, essentially, to expand MR. SCHIFF: I'm sorry, when you met Bannon, had he already had an affiliation with SCL? MR. WYLIE: No. He had several meetings in the United States that I wasn't privy to. MR. SCHIFF: So this was when Nix was trying to establish -MR. WYLIE: Yeah, so Alexander Nix first started meeting with him after an introduction from a Page 11 1 2 mutual contact. When Steve Bannon first came over to the 3 United States and met with me, it was -- he was -- I 4 was just told to go and speak with him, and he just had 5 lots of questions, namely about research into sort of 6 culture, and what is culture, and has culture changed. 7 The research that I was looking at was in whether we 8 can take data sets from different sources and predict 9 psychological disposition. 10 One of the things I spoke to him about was if 11 you -- there is something called the psycholexical 12 hypothesis, which is that there is meaning and 13 information encoded in language. 14 at how different languages describe certain things, and 15 the different kinds of adjectives that we use, there 16 may be a latent construct there. 17 And so when you look And so, just to translate that into something 18 more tangible, the words that we use to describe 19 ourselves as people, in terms of personality traits, 20 are often the same words that we would use to describe 21 cultures sort of more generally. 22 indulge a stereotype for example, what are Italians So if you just Page 12 1 like? 2 more extroverted. 3 be more organized, more diligent, more rule-oriented, 4 right? 5 construct there that could be measured and engaged in 6 some way. 7 They might be more passionate, they might be What are Germans like? They might And that there might be some kind of latent Because I said to him, I said, "If you are 8 interested in culture change, you first have to define 9 what culture is, and -- because if you're -- if you 10 just have this sort of amorphous construct, you can't 11 actually do anything with it." 12 appealing to him because the research that we were 13 doing in terms of profiling people in terms of 14 personality traits likely could extend to culture, 15 because culture is just an aggregation of people. 16 you think about what are the units of culture, the 17 units of culture are people. 18 So that was quite If And so he continued having meetings with 19 Alexander Nix. He very much liked the idea of working 20 with a military contractor, because when he talks about 21 culture wars, I think he literally intends those words 22 in the sense that if you want to fight a war you need Page 13 1 to build an arsenal of weaponry. 2 and contact than a military contractor? 3 And who better to go So, after having sort of several months of 4 meetings, I got introduced -- I got flown over to New 5 York, and I got introduced to Robert Mercer, Rebekah 6 Mercer. 7 8 Steve Bannon also was at this meeting. MR. SCHIFF: And -- So this would have been early 2014? 9 MR. WYLIE: 10 the end of November. 11 quite quickly. 12 met Steve Bannon. He -- we had several meetings and 13 conference calls. And then, sort of mid to the end of 14 November, that's when I went to New York to meet with 15 the Mercers. 16 This would have been -- we're now So by -- so the -- it happened So there was -- October is when I first So Alexander Nix joined me at that meeting, 17 Steve Bannon was there, Robert Mercer was there, 18 Rebekah Mercer was there. 19 essentially, we just spoke about the research that we 20 were doing. 21 22 And that meeting -- One of the sort of impressions that I got was that -- Page 14 1 2 MR. SCHIFF: Before you tell me your impression, if I could -- 3 MR. WYLIE: 4 MR. SCHIFF: Yeah? -- what do you know of the 5 financing of SCL before the Mercers entered the 6 picture? 7 Did they have any patrons like the Mercers? 8 Did they have -- 9 MR. WYLIE: No. At the -- they weren't -- it 10 wasn't -- they didn't have any sort of patrons or sort 11 of wealthy backers at that level, at that sort of 12 billionaire level. 13 money through just predominantly contracts with 14 military clients. 15 So they were going -- they made MR. SCHIFF: You know, there has been 16 speculation and sometimes reporting that SCL may have 17 received Russian financial backing through oligarchs or 18 otherwise. 19 either in the period before the Mercers came along -- Did that ever come to your attention, 20 MR. WYLIE: 21 MR. SCHIFF: 22 MR. WYLIE: So --- or after? At the time I wasn't aware of Page 15 1 that. 2 Tchenguiz, who was one of the shareholders at the time, 3 had quite a few Russian contacts, business partners, 4 that sort of thing. 5 was set up, he was removed as a shareholder, as I 6 understand it. 7 but the impression that I got was that the lawyer -- 8 the Mercer lawyers were concerned about some of his 9 previous business dealings. Although I have heard that, for example, Vincent I don't know the particulars of that, 10 MR. SCHIFF: 11 Mercers coming along? 12 Although when Cambridge Analytica MR. WYLIE: So he was a investor prior to the I don't know if I would -- I don't 13 know if he was an investor or merely just a shareholder 14 in some capacity. 15 16 17 18 19 MR. SCHIFF: And can you spell his name for us? MR. WYLIE: I couldn't, sorry. I can get you the spelling of the name. MR. SCHIFF: And was Dmitry Firtash one of his 20 cohorts that -- you mentioned that he -- that you had 21 heard there might be -- 22 MR. WYLIE: Yeah. Page 16 1 MR. SCHIFF: 2 MR. WYLIE: -- business partners of his -Yeah. I -- so to be clear, I've 3 never met Dmitry Firtash, but my understanding was that 4 he was an associate of some -- in some way. 5 wasn't involved in those meetings. 6 MR. SCHIFF: But I So, in terms of foreign 7 investment prior to the Mercers, there is nothing that 8 you have firsthand knowledge about? 9 MR. WYLIE: Not in terms of things that I saw 10 directly, just simply what I was sort of generally told 11 at the time. 12 MR. SCHIFF: Anything that you have heard 13 indirectly that we might investigate further, in terms 14 of Russian finance or -- 15 MR. WYLIE: What I can do is I can make a note 16 of that, and then speak to Tamsin and look through some 17 of the documents. 18 interested in, I can come back with you with an answer 19 on that. If that's something that you're 20 MR. SCHIFF: 21 MR. WYLIE: 22 MR. SCHIFF: Oh, very much so. Yeah. Thank you. Page 17 1 2 MR. WYLIE: I just don't want to misspeak whilst I'm -- 3 MR. SCHIFF: 4 MR. WYLIE: 5 MR. SCHIFF: 6 Yeah. MR. WYLIE: 8 MR. SCHIFF: 10 So I cut you off when you were describing the meeting you had -- 7 9 No, I appreciate that. Sure. -- with the Mercers and Steve Bannon -MR. WYLIE: So the impression that I got 11 originally was that, you know, Robert Mercer, you know, 12 he made his money in -- you know, he was quite 13 technical as a guy. 14 linguistics. 15 this money in algorithms. 16 He has a Ph.D. in computational So he understands algorithms; he made all The impression that I got, or at least what 17 Alexander Nix said, was -- sort of the carrot on the -- 18 you know, on the stick, was if we were able to sort of 19 predict behavior of each individual person in a 20 particular country like the United States more 21 accurately than anybody else, you know, for the 22 financial sector, if you're looking at investments, if Page 18 1 you can predict consumer trends before anyone else 2 could, that would be quite useful. 3 sense that Robert Mercer would be interested in that. 4 And so, it made But it turns out that, you know, he's already 5 a billionaire, so he doesn't need to make more money, 6 and he was more interested in what Steve Bannon was 7 interested in, which is, you know, can you change a 8 culture, and how would you do that, and how would you 9 quantify a culture, how would you quantify people, and 10 what could you do -- what kind of interventions could 11 you do in order to, you know, shift a cohort of people 12 from here to here. 13 So we spoke about some of the research that we 14 were doing as a firm. 15 research wasn't at the time in the United States. 16 because we were looking at fundamental human traits 17 that would be generally applicable anywhere, he decided 18 to invest. 19 Just to be clear, most of that But So a decision was made that, rather than 20 setting up -- rather than acquiring SCL, a new company 21 would be set up, registered in Delaware. 22 Bannon picked the name Cambridge Analytica. So Steve That Page 19 1 company was set up in part because, as I understand it, 2 there are restrictions on what a foreign company and 3 foreign nationals are allowed to do with respect to 4 elections and foreign agent restrictions. 5 So they -- the lawyers were fairly clear that 6 they needed an American entity to operate in the United 7 States. 8 Republican billionaire acquiring a foreign military 9 contractor to then work in American politics probably 10 And also, from an optical perspective, a wouldn't -- someone would notice that. 11 So Cambridge Analytica was set up. 12 was essentially that Robert Mercer, through one of his 13 investment vehicles, would put in $15 million US into 14 that company as start-up capital. 15 division of SCL Group would become a minority 16 shareholder of Cambridge Analytica. 17 would become the majority shareholder of Cambridge 18 Analytica. 19 The deal The elections Robert Mercer SCL would then assign its intellectual 20 property to Cambridge Analytica. Cambridge Analytica, 21 in return, would license back that intellectual 22 property to SCL Group, and then also offer it an Page 20 1 exclusive service provision contract, so that any 2 clients of Cambridge Analytica would have to be 3 serviced by SCL Group using the intellectual property 4 that it sort of cycled through Cambridge Analytica. 5 So, to be clear, Cambridge Analytica, to the 6 best of my knowledge, never had any staff. 7 only had intellectual property rights and it was where 8 the money from Robert Mercer was originally deposited. 9 10 The -- it But the offices and staff were all based out of SCL Group in London. 11 MR. SCHIFF: So what was the point -- I know 12 you were privy to some of the discussions about the 13 organization of all this. 14 MR. WYLIE: 15 MR. SCHIFF: Yeah. What was the point of that rather 16 Byzantine structure, if it was designed to meet 17 American legal requirements in terms of not having 18 foreigners influence decision-making and campaigning? 19 That doesn't seem to overcome that problem, if then 20 those people are actually employed in the United 21 States. 22 back. It's just where the IP is held and licensed Page 21 1 MR. WYLIE: So it wasn't my decision to do 2 that. 3 Steve Bannon, and Robert Mercer to set it up that way. 4 The problem I think that they encountered is It was a decision made between Alexander Nix, 5 that if you are going and acquiring the intellectual 6 property of a company, a large part of servicing that 7 is people who are experienced and know how that 8 actually works, and how to use it. 9 people, you know, were British or European or other 10 11 And all of those foreign nationals. And so, that, you know -- and one of the 12 things that I published was some of the legal advice 13 that Cambridge Analytica got. 14 seen that memo to Steve Bannon and Robert Mercer that 15 -- even the fact that Alexander Nix was the CEO was 16 questionable because he was a foreign national. 17 whole thing was -- 18 19 MR. SCHIFF: So the Was that legal memo written prior to the incorporation of Cambridge Analytica? 20 MR. WYLIE: 21 MR. SCHIFF: 22 I'm not sure if you've No, it was written after. So -- Because I would -- then what drove the -- that rather odd structure? Page 22 1 MR. WYLIE: Right. So my understanding was 2 that it was also in part a way -- so I think -- to be 3 honest, I think it was because Robert Mercer wanted 4 something that was his own, but it couldn't be SCL. 5 The other advantage of having Cambridge 6 Analytica in the United States was that, at least to my 7 knowledge -- so I am not a U.S. elections law or 8 finance expert, so you need to consult an expert on 9 that, but my understanding was that an advantage of 10 having a company that services campaigns is that if you 11 are the shareholder of that company, when you invest 12 money into that company, it does -- it's not counted in 13 elections reporting as a campaign donation, because you 14 are an owner of a company investing more resources or 15 capital into that company. 16 So it created a vehicle whereby you could do a 17 lot of research and development, ultimately for the 18 benefit of your preferred campaigns, but it wasn't a 19 reportable campaign donation because when you're 20 investing in a company you're an investor. 21 22 MR. SCHIFF: I mean that, I imagine, would be true if you have more than one client. If you are Page 23 1 doing this all in the service of the same client, then 2 it would simply be probably an ineffectual effort at 3 circumvention. 4 But it sounds like, if that were the goal, you 5 could do that with a very different structure than 6 Cambridge Analytica. 7 of Cambridge Analytica you would be able to invest in, 8 and therefore subsidize the work that that firm is 9 doing. 10 I mean any form of incorporation I guess what I'm trying to figure out is 11 whether there was some -- what the rationale was for 12 that -- the license-back structure and that particular 13 form of holding of intellectual property. 14 MR. WYLIE: I -- to be honest, I found it 15 confusing, also. 16 what Alexander Nix and Steve Bannon wanted, so that's 17 what the lawyers ended up setting up. And it's something that -- that's 18 It -- I think one of the -- again, one of the 19 problems is that if you -- you could -- if you had only 20 set up Cambridge Analytica as an independent, separate, 21 American company with no license-back, you would then 22 have to staff it. And so the problem with that is that Page 24 1 the -- all the people who knew about it -- it's -- 2 you're essentially buying the IP without any of the 3 brains behind it. 4 And so that, I think, would have created a 5 problem with actually creating a functional company, 6 because normally, when you acquire a company or create 7 a start-up, you know, the foundational team or the 8 experts sort of are part and parcel of that company. 9 MR. SCHIFF: 10 MR. WYLIE: Right, I see. So it would have sort of created a 11 -- there would have been no point in really doing it, 12 because a lot of the value is actually in the people 13 who had experience in the UK. 14 15 MR. SCHIFF: Mr. Quigley? 17 MR. QUIGLEY: 19 Okay. I think my time is up. 16 18 I see. Thanks for being here. And for the record, you -- again, you left SCL when? MR. WYLIE: In November of 2014. I sent in my 20 notice in July, at the end of July 2014, and then I 21 phased out my involvement until November. 22 MR. QUIGLEY: After you left, were you able to Page 25 1 continue to be aware of what was going on within the 2 company? Did you have -- 3 MR. WYLIE: 4 MR. QUIGLEY: 5 Yes. -- people there? Did you have -- 6 MR. WYLIE: 7 MR. QUIGLEY: 8 How much do you know of what was still going 9 Yes. How did that work? on, and how -- 10 MR. WYLIE: Because when I -- when Cambridge 11 Analytica was formed, I was responsible for building 12 out the teams. 13 of the different teams. I was responsible for building out all 14 MR. QUIGLEY: 15 MR. WYLIE: Okay. So I had very good working 16 relationships with most people there, because I 17 previously worked on projects. 18 19 MR. QUIGLEY: And we're talking about how many people? 20 MR. WYLIE: 21 around 40 to 50, give or take. 22 the exact number, if that's of interest. By the time I left, probably I can provide you with Page 26 1 2 MR. QUIGLEY: And approximately how many of those 40 to 50 did you stay in touch with? 3 MR. WYLIE: Give or take, 15, 15 or so. But 4 they were the sort of the leadership tier of each of 5 the teams. 6 MR. QUIGLEY: 7 MR. WYLIE: How many teams were there? You had data science, you had 8 psychology, creative and messaging, operations. 9 six or seven. 10 11 MR. QUIGLEY: them? 12 13 Email? MR. WYLIE: I would just see -- I mean it was based in London and I was living in London, so I -MR. QUIGLEY: 15 MR. WYLIE: 17 18 19 How did you stay in touch with Social media? 14 16 Around You saw them. I see them all the time. Yeah, yeah. And then they would send me updates, or forward me emails, or tell me what's happening. MR. QUIGLEY: And that relationship, those 20 relationships, that communication to this day continues 21 in your -- 22 MR. WYLIE: Continued until -- Page 27 1 2 3 MR. QUIGLEY: Obviously, some of it got hampered -MR. WYLIE: Until they all started leaving. 4 The problem for a lot of them was that when they joined 5 the company -- after Steve Bannon took over, the 6 company took a much more hard line on the types of 7 research that we were going to do, particularly in the 8 United States. 9 MR. QUIGLEY: 10 MR. WYLIE: 11 MR. QUIGLEY: 12 MR. WYLIE: 13 MR. QUIGLEY: 14 MR. WYLIE: Cambridge, or -- Cambridge Analytica. Right. It's sort of difficult to -That's fine. So, to be clear, all the staff 15 were technically SCL staff, because CA didn't have any 16 staff, but operating under the brand or the name 17 Cambridge Analytica. 18 19 MR. QUIGLEY: understand. So the staff were all getting paid by SCL. 20 MR. WYLIE: 21 MR. QUIGLEY: 22 Well, I just -- I want to Yes. But they were -- how many -- what percentage were just doing Cambridge stuff? Page 28 1 2 MR. WYLIE: Most, if not all, because -- I mean it's very hard to distinguish between the two. 3 MR. QUIGLEY: 4 MR. WYLIE: Right. Because, you know, if you are 5 speaking to an American client, it's -- you put on your 6 Cambridge Analytica hat. 7 Cambridge Analytica email. 8 you know, a Brit or foreign -- other foreign client, 9 you might be speaking to them as SCL Group. 10 11 MR. QUIGLEY: MR. WYLIE: 13 MR. QUIGLEY: 15 But if you're speaking to, So you've heard the expression "shell company," right? 12 14 You know, you use your Yes. Would you refer to Cambridge as that? MR. WYLIE: In a -- so I'm not sure if that 16 word has a specifically legal meaning in the United 17 States, but in the sort of general parlance sense -- 18 MR. QUIGLEY: 19 MR. WYLIE: 20 21 22 Right. -- yes, I think that would be a completely fair assessment. MR. QUIGLEY: How many of the people that you are staying in touch with knew what was going on with Page 29 1 the Trump campaign? 2 MR. WYLIE: 3 MR. QUIGLEY: 4 MR. WYLIE: A couple of them. Okay. The company first started engaging 5 what later became Donald Trump's campaign in the spring 6 of 2015. So actually, very shortly after I left. 7 MR. QUIGLEY: 8 MR. WYLIE: 9 Okay. And I have that confirmed in writing, actually, from SCL's lawyers. 10 So despite the fact that -- despite the Trump 11 campaign saying that they -- the company, you know, 12 only started engaging with them in August 2016, they 13 actually started meeting with Corey Lewandowski, who 14 then later became the campaign manager for Donald 15 Trump, before Donald Trump had even announced that he 16 was considering running for President. 17 18 19 20 MR. QUIGLEY: What date would that be, roughly? MR. WYLIE: It would have -- it was the spring of 2015, and I can get you the exact date in writing. 21 MR. QUIGLEY: 22 MR. WYLIE: The spring of what, 2015? The -- 2015, yeah. Page 30 1 2 MR. QUIGLEY: And so you understood that the Cambridge staff was meeting with Corey Lewandowski. 3 MR. WYLIE: Yes, and I have it confirmed from 4 SCL's own lawyers that they were. 5 MR. QUIGLEY: And when was there, to your 6 understanding, a formal agreement that they would work 7 together? 8 MR. WYLIE: That I don't have. 9 MR. QUIGLEY: And you say that there is a 10 couple that you stayed in touch with that were, you 11 know, working on the Trump -- or working with the Trump 12 campaign? 13 MR. WYLIE: 14 MR. QUIGLEY: 15 16 That -Or embedded? You tell me what the -MR. WYLIE: Who were around during the 17 beginnings of the work for the Trump campaign. 18 then left because they found that uncomfortable. 19 20 21 22 MR. QUIGLEY: when they left? How -- and do you know roughly You know, what time frame? MR. WYLIE: can probably -- They I don't have that exact date. I Page 31 1 MR. QUIGLEY: 2 MR. WYLIE: 3 MR. QUIGLEY: 4 Okay. -- get that for you. Sure, yeah. Did they tell you why they were uncomfortable? 5 Or I forgot the word. 6 MR. WYLIE: Is that right? Yeah, uncomfortable, yes, because 7 the -- most of the staff who originally joined 8 Cambridge Analytica were not conservatively-minded 9 people. 10 11 12 13 MR. QUIGLEY: They didn't like the President's -- the candidate's policies. MR. WYLIE: Not just his policies, because Donald Trump is a very particular kind of -- 14 MR. QUIGLEY: 15 MR. WYLIE: 16 MR. QUIGLEY: 17 MR. WYLIE: Okay. -- candidate. So -- He wasn't just a Republican -He wasn't just -- there is a big 18 difference between working for, you know, a Jeb Bush 19 type Republican and a Donald Trump type Republican. 20 MR. QUIGLEY: When we talk about this in 21 reference to -- and my colleagues will get into more 22 detail -- the issues as related to the Facebook data Page 32 1 and how that was used by Cambridge and the Trump -- 2 were any of them making you aware of that before they 3 were involved in the Trump campaign? 4 5 Or was that prior to that information, or that data being switched over? 6 MR. WYLIE: With respect -- so your -- if I 7 can just rephrase your question, your question is about 8 whether or not the Trump campaign was using 9 misappropriated Facebook data. 10 11 MR. QUIGLEY: Right. Did they know? Did you know? 12 MR. WYLIE: They didn't -- I can't confirm 13 that, but I can explain how it would be technically 14 very difficult to achieve anything that they do without 15 that data, given the timeframe that they say they did 16 the work. 17 18 19 MR. QUIGLEY: Yes, can you walk us through that? MR. WYLIE: So the Facebook data -- so maybe I 20 will just quickly explain how that project got set up, 21 and then what the utility of it was. 22 So in the spring of 2014, after the investment Page 33 1 was made into Cambridge Analytica, the next step in the 2 project was to figure out how to acquire a scaled data 3 set on personally identifying Americans. 4 means it's addressable data. 5 some kind of thing that I can match to the electoral 6 register, and that is also useful for modeling very 7 nuanced human constructs, which are very difficult to 8 model. 9 data, and there had to be a lot of it. 10 So -- which It has their name, it has So it had to be sort of a high-quality type of When we started working with several of the 11 professors at the Psychometrics Centre at the 12 University of Cambridge, they had access to 13 applications that were authorized by Facebook that 14 allowed them to collect not just the data on the user 15 of that app, but also the data of the friends of that 16 app. 17 MR. QUIGLEY: 18 didn't approach you? 19 MR. WYLIE: 20 And you approached them? They The Cambridge folks. It was a mutual introduction. So it's difficult to say who approached whom, because -- 21 MR. QUIGLEY: 22 MR. WYLIE: Okay. -- we had people who previously Page 34 1 were at Cambridge who did their psychology degrees 2 there, and so it was just sort of a mutual 3 introduction. 4 5 But -- MR. SCHIFF: Can I interrupt you for just one second? 6 MR. WYLIE: 7 MR. SCHIFF: 8 MR. WYLIE: 9 MR. SCHIFF: Sure. Just for clarification. Yeah. At the point that you're trying 10 to get a data set to move forward on the culture- 11 changing project of Bannon and Mercer's, this takes 12 place after Cambridge Analytica is formed? 13 not being done by SCL, pre-Cambridge Analytica? 14 MR. WYLIE: So this is There was some preliminary 15 research that SCL did prior to Cambridge Analytica 16 being formed in 2013, which was designed to be sort of 17 a bit of an exploratory project in the United States, 18 because the company hadn't, to my knowledge, hadn't 19 worked in the United States before. 20 validation or a sort of -- a proof of concept, or at 21 least showing promise of methods. 22 Virginia, in Fairfax County and Tidewater, the And also, a That was done in Page 35 1 2 Tidewater region of Virginia. So there was data collected there, and various 3 sort of pilots of different experiments that was done 4 that Robert Mercer used to sort of base some of his -- 5 because he wanted to see, you know, could you do this 6 kind of -- you can -- you know, sure, you might be able 7 to do this in country X or Y, but does it -- can you do 8 it in the United States? 9 So -- but to my knowledge, that wasn't 10 necessarily done for a -- any particular campaign, 11 although I do recall meeting with various Republican 12 affiliated campaigners who were interested in it, but 13 they weren't the client, if that makes sense. 14 sort of just a general R&D type project. 15 It was Just to get back to the Facebook data, so the 16 benefit of the application that Kogan offered was that 17 we could replicate existing peer-reviewed methods that 18 the Psychometrics Centre had sort of validated in their 19 research, and published that in journals. 20 sort of best practice or scientific perspective, we are 21 replicating what has already been validated in research 22 and published in journals. So, from a Page 36 1 So that made sense, and also the fact that 2 what he offered allowed a very rapid scaling of a data 3 set because, rather than collecting data one by one, 4 each time somebody would join that app the app would 5 give you access to, you know, 200, 300 other people who 6 were friends with them on Facebook. 7 imagine how much quicker -- you know, it scales, you 8 know, at 300 times faster than if you were to do it 1 9 by 1 by 1 by 1. 10 So then you can So SCL -- and it was SCL that technically paid 11 for that project -- put in around 800 to $900,000 US, 12 just on that data harvesting, which then -- and it then 13 transferred the rights to that to Cambridge Analytica. 14 That data then formed the basis of the algorithms for 15 every single algorithm that was developed at the 16 company. 17 So the reason why that's important is because 18 they may say we didn't use Facebook data on the Trump 19 campaign, but if the algorithms that you have built for 20 your company were all built using Facebook data, it's 21 -- it becomes a slightly semantic difference because 22 you don't actually -- when it comes to targeting Page 37 1 people, you don't actually use the Facebook data 2 directly, you use it to build an algorithm that then 3 creates your target universe of people. 4 So, it's like saying I don't use petrol, I 5 just drive a car. 6 Facebook data, you are indirectly using petrol 7 gasoline. 8 9 Well, if the car is powered by If that makes sense. So the thing that I find difficult to believe, without -- because, the thing is, the company has never 10 offered any evidence that they used any other data 11 sets, right? And it would have been very simple for 12 them to do. They've been asked by the British inquiry 13 to provide that information, but Alexander Nix decided 14 not to show up to his hearing. 15 When you look at the testimony of Brittney 16 Kaiser, who was at the testimony at the British inquiry 17 last week -- Brittney Kaiser was the business 18 development director -- she makes reference to seeing 19 emails about the use of Facebook data as recently as 20 last year. 21 with is Brittney Kaiser. 22 So someone you might want to actually speak MR. QUIGLEY: Does that -Yes. In the end, to the extent Page 38 1 that you know, were any of the people you followed 2 embedded in the Trump campaign later on? 3 4 5 6 7 8 9 10 MR. WYLIE: They weren't embedded in the sense that they went to the United States. MR. QUIGLEY: They were -- The San Antonio operation we hear about. MR. WYLIE: The -- yeah, Project Alamo, or whatever it was called. MR. QUIGLEY: MR. WYLIE: They were based in London. Okay. But I can follow up with you on 11 that, because I can also -- I can give you a list of 12 names of people who might be helpful with respect to 13 that particular -- 14 15 MR. QUIGLEY: It's hard for us -- you were there, and you know these folks, and so -- 16 MR. WYLIE: 17 MR. QUIGLEY: Yeah. In my short time that I have 18 left, ultimately what is of interest to us is Cambridge 19 Analytical people must have known that this information 20 was inappropriately gained, and it was going to be 21 used, and then that fact was going to be used for -- to 22 help the Trump campaign. Page 39 1 2 Can we know, or do we know if the Trump campaign personnel were -- was aware of that fact? 3 MR. WYLIE: This is something that will be 4 discoverable in the next few months, with the inquiries 5 in Britain and the investigation in Britain. 6 can't offer you a definitive answer specifically to the 7 Trump campaign, with respect to the use of Facebook 8 data. 9 So I But what I can tell you is how difficult it 10 would be. 11 the algorithms that were developed using that -- 12 13 MR. QUIGLEY: Trump? You're referring to the Trump campaign? 14 15 They used -- and in their own words -- used MR. WYLIE: Sorry, I am referring to Cambridge Analytica. 16 MR. QUIGLEY: 17 MR. WYLIE: Okay. Cambridge Analytica used 18 algorithms, the algorithms that I am talking about that 19 were developed using the misappropriated Facebook data, 20 up until the summer of 2016. 21 stopped using it. 22 They then claim that they The problem with that claim is that, in order Page 40 1 to develop what they developed, it took several million 2 dollars and months and months and months and months and 3 months and months of work. 4 had a house that we built, and then we knocked it down 5 and the next day we rebuilt it in a day. 6 technical perspective, their explanation does not make 7 sense. 8 9 And so it's like saying we It -- from a It also doesn't make sense because the data sets that they claim to have used on the Trump 10 campaign, the commercial data sets that they claimed to 11 use on the Trump campaign that replaced the Facebook 12 data, one of the things that I can provide is internal 13 assessments of those data sets that show that those 14 data sets weren't sufficient to accurately model the 15 concepts that they -- 16 MR. QUIGLEY: So let me try to summarize real 17 quick. The Cambridge Analytical people had to know 18 what they were using was information at least derived 19 from the original -- 20 MR. WYLIE: 21 MR. QUIGLEY: 22 MR. WYLIE: Yeah. -- poisonous tree. Yeah. Page 41 1 MR. QUIGLEY: And it's virtually impossible 2 that the information that was ultimately used in the 3 Alamo operation, or whatever, wasn't derived from the 4 same source. 5 6 MR. WYLIE: make is that -- 7 MR. QUIGLEY: 8 MR. WYLIE: 9 10 11 Well, the other point that I would Would you say that's correct? Yeah. Well, I would -- I find it -- they have not yet offered an explanation -MR. QUIGLEY: Of where any of the other data could have come from. 12 MR. WYLIE: 13 MR. QUIGLEY: 14 MR. WYLIE: -- how they did that. Right. Because that is, in my view, 15 pulling off a miracle. 16 -- you know, that's very impressive. 17 MR. QUIGLEY: 18 source for the data. 19 MR. WYLIE: If they did that, they are the And they didn't have another No. This is the problem with 20 their explanation, is that -- one of the things that I 21 can show you is that the data that they claim to use 22 doesn't work for what they claimed they did for the Page 42 1 Trump campaign by -- so it doesn't make sense that you 2 replaced the Facebook-derived information with this 3 consumer data set, because I know that that consumer 4 data set doesn't work for what you claim it does. 5 MR. QUIGLEY: And finally, the people who are 6 embedded in Texas would have had to at least known 7 this, and that you're suggesting that in the coming 8 short period of time the inquiry in Great Britain will 9 detail this and let us know more. 10 MR. WYLIE: The parliamentary -- so the 11 parliamentary inquiry is going to demand Alexander 12 Nix's presence if he continues to refuse to -- if he 13 keeps skipping his hearings. 14 15 In addition, there is an ongoing investigation led by the information commissioner's office in the UK. 16 There is also an element of that investigation with 17 the National Crime Agency, which is the British FBI. 18 And if there is relevant information for the American 19 authorities, my understanding is that the National 20 Crime Agency will be coordinating with the FBI and 21 Department of Justice on those. 22 MR. SCHIFF: Thank you, Mr. Quigley. I just Page 43 1 have a couple of follow-up questions and then I'm going 2 to pass to my colleague, Mr. Himes. 3 MR. WYLIE: 4 MR. SCHIFF: Sure. You mentioned that Cambridge 5 Analytica, or SCL, paid $800,000 or $900,000 to acquire 6 one of the early data sets from Kogan. 7 correct? 8 9 MR. WYLIE: Is that Those -- so what I can provide to you is the invoices and payment transactions that were 10 made to the platforms that were recruiting participants 11 to join the apps. 12 the apps, and how much did that cost. 13 up all the invoices, it's between 800 and $900,000. 14 15 MR. SCHIFF: So was Kogan the developer of the MR. WYLIE: Facebook was the developer of the MR. SCHIFF: And what was Kogan's role, then, app. 18 19 And if you add apps that -- 16 17 So it details how many people joined vis a vis the apps and the -- 20 MR. WYLIE: 21 correct myself. 22 API. So all Facebook -- well, I should Facebook provides what's called the That's like a bridge. And so it sort of is the Page 44 1 bridge that allows you to access the system. 2 to develop sort of the interface of the app, as it 3 were. 4 the app; Facebook was the authorizer of the app. 5 You have So in that respect, Kogan was the developer of The role that Kogan offered was twofold. 6 Firstly, he had the rights to the app that had these 7 permissions that Facebook rented. 8 expert in psychometrics. 9 lot of the psychological surveys that were then used on Secondly, he was an And so he also developed a 10 the app to -- for people to answer, so that -- which 11 became the -- what's called the training set for the 12 algorithms. 13 So that data, the answers from the surveys 14 would be related to that person's Facebook data, and 15 then relationships between how they answered and what 16 they liked on Facebook would then develop the algorithm 17 to them profile all of their friends. 18 MR. SCHIFF: So what's the nature of the 19 service that Kogan was providing to Cambridge SCL, in 20 terms of the acquisition of the data from Facebook? 21 It -- 22 MR. WYLIE: He sort of -- he had the app. So Page 45 1 he offered the use of the app. 2 MR. SCHIFF: 3 So were there payments to him, essentially, like a consultant -- 4 MR. WYLIE: 5 MR. SCHIFF: So --- and then payments to the 6 intermediary companies that actually did the physical 7 harvesting of the data on Facebook? 8 9 MR. WYLIE: So, originally, Kogan said -- and I believed him at the time -- that he was interested in 10 working on this data acquisition project because it 11 would have given his team at Cambridge access to 12 funding that he couldn't otherwise have got through an 13 academic grant with almost no sort of strings attached 14 or reporting rules, or all of that. 15 take that data set that was then created and set up an 16 academic institute with it to sort of explore this 17 emerging field called computational psychology, or 18 computational sociology, which is sort of blending 19 computer science with psychology. 20 And he wanted to So, the original agreement that SCL/CA had 21 with Kogan was that he wouldn't actually get -- he 22 wouldn't make any profits. He would pass the invoices Page 46 1 from the recruitment platforms that were used -- so 2 that was Amazon, MTurk, and a company called Quatrics 3 -- and that SCL would then pay those invoices on his 4 behalf. 5 So what I can provide to the committee is the 6 -- some of the payment transactions that showed that 7 SCL was just paying for them directly. 8 to me at the time, because it -- his motivation seems 9 to be setting up this institute, so it wasn't setting 10 up a company, it wasn't, you know, going out and doing 11 anything with it, it was to do academic research. 12 That made sense One of the things that Facebook told me after 13 the fact was that they were informed by Kogan that his 14 activities were also for academic purposes. 15 was my understanding, and that's what Facebook told me 16 was also their understanding. 17 about that. 18 So that You need to ask Facebook However, after I left Cambridge Analytica, 19 Kogan went back to Alexander Nix and decided to 20 negotiate a more commercial contract. 21 apparent that he was interested not just, I guess, in 22 this academic institute, but then also commercializing And it became Page 47 1 that data and further deepening his relationship with 2 Cambridge Analytica. 3 And so, the institute that he talked about -- 4 and I can provide you with emails that talk about sort 5 of conceptualization -- actually never got set up. 6 I'm not sure exactly what happened there, but he went 7 down a path that was quite different than what 8 originally in my understanding was. 9 MR. SCHIFF: So Now, you're probably more 10 familiar with this than I am; I think I've only seen a 11 summary of what Aleksandr Kogan has had to say. 12 MR. WYLIE: 13 MR. SCHIFF: 14 Yes. He takes issue with the idea that he provided this data to Cambridge Analytica SCL. 15 MR. WYLIE: 16 MR. SCHIFF: Right. Is he being hyper-technical about 17 this, in the sense that it was -- the invoices went to 18 other entities and he was just the facilitator, or -- 19 MR. WYLIE: 20 MR. SCHIFF: 21 MR. WYLIE: 22 I -How do you explain that? I take issue with a lot of his statements because, for me, I feel like he is being, as Page 48 1 you said, slightly overly technical with his 2 interpretation of the situation. 3 You know, when you, for example, transform 4 data -- so you squish some data together or you split 5 some data apart, rename it, whatever, that is called a 6 derivative. 7 and I merge them into one column called gender, that is 8 a derivative of those two columns. 9 conveyed is the same. 10 So if I take two columns, male and female, The information So, one of the things that he provided is 11 derivative information from Facebook data. 12 takes the position that that's not Facebook data, 13 that's a derivative and, therefore, it's not prepped to 14 say Facebook data. 15 And so he But one of the things that I can provide to 16 you is the explanation of the process, which is what he 17 explained to the company at the time, which is -- makes 18 very clear that this is Facebook-origin data and it's 19 being processed in this way, and this is what you get 20 after we've processed it in this particular way to make 21 it easier for you to use. 22 MR. SCHIFF: And last question. At the time Page 49 1 that Cambridge is working with Kogan, does this -- does 2 it predate Bannon, or is it all post-dating Bannon? 3 Basically, is the relationship either done or 4 solidified with Kogan for the very purpose of meeting 5 the kind of Mercer-Bannon idea of -- 6 MR. WYLIE: 7 MR. SCHIFF: 8 9 Yeah. -- gathering this data and being able to apply it in their culture war? MR. WYLIE: Yeah. So we started the project 10 with Kogan after the Mercers invested in Cambridge 11 Analytica and after Steve Bannon took over. 12 and the point of the project was because after they set 13 up Cambridge Analytica, the Mercers and Steve Bannon, 14 they then came to us and said we need to have 15 everything ready, essentially, by September 2014, 16 which, to me, seemed sort of outrageously ambitious. 17 So it -- And then -- but it was actually because Steve 18 wanted to actually try it out in the 2014 mid-terms. 19 So he wanted to be able to pass on this to campaigns to 20 use, which meant that the deadline was essentially 21 September. 22 that it scaled -- I mean we were able to acquire that And so, the benefit of Kogan's process was Page 50 1 data set in sort of two-and-a-half months. 2 how quick it was. 3 So that's And so, from Steve Bannon's perspective, that 4 was ideal because he got what he wanted. 5 massive data set in the time frame that would allow him 6 to then apply these algorithms targeting for the 2014 7 mid-terms. He got this 8 MR. SCHIFF: 9 Mr. Himes? 10 MR. HIMES: 11 it's a big question. 12 away from the chronology, which is important. 13 would be helpful to me and, I suspect, to the committee 14 if you could tell us a little bit more about the 15 concepts that we're talking about here. Thank you. Thank you. Just one question, but And I'm going to ask you to step But it 16 MR. WYLIE: Sure. 17 MR. HIMES: Psychometrics, computational 18 psychology, micro-targeting -- I have been swimming in 19 those terms for a long time and I am not sure I fully 20 understand them. So a three-part question. 21 One is I've got a personal Facebook page, I 22 follow my college roommates around, I like a piece of Page 51 1 music and a pair of shoes every once in a while. 2 MR. WYLIE: Yeah. 3 MR. HIMES: In terms that that a guy can 4 understand, what is the aspiration here, in terms of 5 what the Trump campaign was trying to do? 6 Can you just walk us through an example about 7 how you take my data, use micro-targeting 8 psychometrics, et cetera, to achieve some end? 9 just walk us through that example? 10 Can you That's question -- part one of the question. 11 Part two is what do you think was actually 12 done? In other words, you made an interesting comment 13 earlier, which was that there was a different set of 14 data which could not be used to actually achieve the 15 goal that was -- 16 MR. WYLIE: Yeah. 17 MR. HIMES: That's part two. 18 Then part three is Dr. Kogan has questioned 19 the effectiveness of any of this. 20 this, in the context of elections, not burning down a 21 new restaurant in Sri Lanka -- 22 MR. WYLIE: Yeah, yeah. In part three is Page 52 1 MR. HIMES: -- which we read about, but in the 2 context of elections, is there evidence that this is 3 effective? 4 In other words, are we actually looking into 5 whether election law was violated, or are we actually 6 looking into whether the election was altered through 7 this mechanism? 8 9 MR. WYLIE: first question. Sure. So I will start with your So let me just quickly define some of 10 the things that I've talked about, just so that you -- 11 so psychometrics is the measurement of psychological 12 constructs. 13 because human psychology is not directly measurable, 14 like different chemicals in your blood or in your 15 nervous system, where you can literally measure the 16 exact amounts of something. 17 study of the best way to measure discrete constructs 18 like personality traits, for example. 19 The reason why it's an area of research is So psychometrics is the Micro-targeting is where you take data, a 20 baseline data set on voters -- whether that's consumer 21 data, social data, et cetera -- and look for 22 relationships between that data and what's called a Page 53 1 target variable. 2 3 A target variable is anything that you want to predict for. So -- 4 MR. HIMES: On the -- 5 MR. WYLIE: I mean, yes, it's a -- yeah. 6 MR. HIMES: Okay. 7 MR. WYLIE: So -- and then, to your question 8 about how is Facebook data useful, when you look 9 particularly at human psychological constructs like 10 personality traits, right, certain kinds of information 11 give you better insight or information. 12 So when you look at consumer data sets, 13 knowing that somebody shops at Wal-Mart, for example, 14 isn't very indicative of who they are as a person, 15 because it might be that they are close to a Wal-Mart, 16 or it might be that their socio-economic status, you 17 know, dictates that they shop at, you know, a discount 18 store. 19 Wal-Mart because you're this type of person, 20 psychologically. 21 22 So it doesn't mean that, you know, you shop at So the advantage of social data -- so Facebook data or Twitter data, for example, or Instagram data, Page 54 1 or what's called click stream data, which is sort of 2 your browsing history -- is that you are curating your 3 identity in a way that is much more insightful for the 4 constructs, for the psychological constructs that we 5 want to measure. 6 So, if you just take a step back and just 7 imagine that you're on a date, for example, a first 8 date, the questions that you ask -- you know, what kind 9 of music do you listen to, what kind of movies do you 10 watch, you know, what do you like to do for fun -- the 11 reasons you ask those questions is because you are 12 developing, you know, an intuition as to who is this 13 person. 14 And when you think about the things that 15 you're liking on Facebook, even though they sort of 16 seem innocuous when you're liking them individually -- 17 you're liking this particular brand of clothing, or 18 you're liking this particular music band -- when you 19 look at it collectively, actually you're creating a 20 profile of who you are, as a person. 21 22 And so the advantage of Facebook data is that the types of information that you're revealing about Page 55 1 yourself is indeed strongly related to personality 2 traits. 3 place, which makes it really easy to acquire, because I 4 just need to look at your likes page, and it's all in 5 one place. 6 just looking at that. And secondly, you're putting it all in one And then I can learn a lot about you from 7 Does that -- 8 MR. HIMES: Yes, thank you. 9 MR. WYLIE: -- answer your question? 10 MR. HIMES: I think part two, which is what I 11 hope you're getting to, is what is the objective, 12 therefore, and the promise to somebody like the Trump 13 campaign? 14 that I can change your behavior, or both? 15 That I can identify who you are, or is it MR. WYLIE: Both, because in order to change 16 your behavior, you first have to identify how that 17 person thinks and engages. 18 So if I'm going -- so, for example, if you 19 look at different traits like openness, for example. 20 So higher degrees of openness -- that's things like 21 curiosity, novelty-seeking, tolerance for ambiguity, 22 these types of things -- correlated to liberal views, Page 56 1 which makes sense, because if you're open to new ideas, 2 that generally means that you're more liberal. 3 The inverse, where you're not, you're less 4 curious, you don't like new things, you're more likely 5 to be a conservative. 6 Knowing the level of that trait helps me be 7 able to speak to you, because, on one hand, somebody 8 who is higher on openness might be enticed by something 9 new, so I might use language about how this is new, or 10 this is changing, or this is revolutionary. 11 somebody is lower in openness, I wouldn't use that 12 language. 13 If So it helps you identify how a person engages 14 with information and then, secondly, if you can 15 identify how they engage with information, you can 16 better curate your message to that person. 17 Perhaps a more tangible example in politics is 18 when you think about something, sort of the most 19 generic political message, jobs in the economy -- you 20 hear a lot of politicians talk about jobs in the 21 economy. 22 say it's important. And when you look at polling, lots of people But one of the things that we Page 57 1 found is that, actually, it's a socially rehearsed 2 answer. 3 important because they think it should be, but that 4 doesn't mean that they actually care, because they 5 already have a job. 6 So people say that jobs and the economy is And so, when we start to sort of unpack what a 7 job is, one of the things that we found was that 8 actually, if we cite jobs don't matter, you know, 9 achievement matters, it's -- the feeling that 10 achievement that I get matters, or the sense of 11 security matters, or the respect that I get from my 12 children when I come home matters, that resonated a lot 13 more with people, particularly when we matched it to 14 particular traits that meant that, you know, you were 15 higher in achievements driving, or you were more 16 anxiety-prone, so you liked things about feeling safe 17 and secure. If that -- does that kind of make sense? 18 So it's about translating abstract constructs 19 or concepts like jobs and the economy, which sort of -- 20 people kind -- which sort of goes over people's heads, 21 into something that actually resonates with who you 22 are. But who you are is different from this person and Page 58 1 this person and this person. 2 So I might translate why a job is important to 3 you differently than to this person than to that 4 person. 5 6 Does that -MR. HIMES: It does, yeah. And if you could just sort of move on to -- 7 MR. WYLIE: Yeah. 8 MR. HIMES: That was quite helpful. 9 10 So what was -- what's your understanding of what was actually done by the campaign? 11 MR. WYLIE: Right. 12 MR. HIMES: And my last question was do you -- 13 is there a reason to believe that this was effective 14 and actually changed -- 15 MR. WYLIE: Yeah. 16 MR. HIMES: -- people's voting behavior? 17 MR. WYLIE: So I very strongly disagree with 18 Kogan's assessment of the effectiveness of profiling. 19 And the reason I do is because, first of all, I've seen 20 it work, myself. 21 word for it, I can provide you with copious amounts of 22 peer-reviewed academic research in top psychology Secondly, you don't have to take my Page 59 1 journals and top data science journals that show that 2 this works. 3 value. 4 using Facebook data, Facebook data would, therefore, be 5 useless, and the company would have no value. 6 7 Does that statement apply to the electoral context? I mean I got it when I'm buying shoes and that sort of thing, but the electoral context? 10 11 If you could not target people effectively MR. HIMES: 8 9 If it didn't work, Facebook wouldn't have MR. WYLIE: Sure. But it is -- yes. Yes, it does. 12 MS. SPEIER: 13 MR. WYLIE: 14 MS. SPEIER: 15 MR. WYLIE: Proof? Sorry? What kind of proof do you have? I can send you papers, if you 16 would like. 17 follow up with documentation if that is -- if you 18 prefer. 19 20 I can explain more about it, or I can MR. HIMES: I think that would -- we have limited time, so -- 21 MR. WYLIE: Yeah. 22 MR. HIMES: And I really appreciate the Page 60 1 background. 2 So that would be great. I asked a group of experts who happened to be 3 presenting as to -- the same question, and they all 4 sort of said, well, this is scary stuff. 5 electoral context there is not a lot of peer-reviewed 6 data that -- peer-reviewed conclusions that suggest 7 that it actually does affect the -- and I asked U.S. 8 electoral. 9 In the But I guess, again, I don't want to -- we need 10 to get back to the chronology. 11 MR. WYLIE: Sure. 12 MR. HIMES: The last piece -- so yes, the 13 answer is please do, if you wouldn't mind sending us -- 14 MR. WYLIE: I am happy to. 15 MR. HIMES: -- whatever you have that is about 16 17 the electoral context. But I don't want to lose this piece of -- you 18 said something very interesting, which was that the 19 data that they had didn't allow them to do what they 20 said they would do for the campaign. 21 22 MR. WYLIE: Right. So the data -- the commercial data sets that they had -- so, first of all, Page 61 1 when Alexander Nix says 4,000 to 5,000 "data points," 2 what he is actually referring to is actually the 3 feature set. 4 Feature set, if you just think about it simply 5 as, like, column titles -- so what he's saying is there 6 is there is 4,000 to 5,000 column titles. 7 mean that those are filled with data. 8 9 That doesn't So when you look at what's called the coverage rate, so that's, in this column, how much -- how many 10 records are filled with an observation for that column. 11 You have some features that have 2 percent, 5 percent, 12 you know, 10 percent, which means that you actually 13 have something called a highly sparse matrix, which 14 means that it's -- you have a lot of possible features, 15 but it's mostly empty. 16 And what that means is that it actually isn't 17 very helpful, because if you only have information on 18 two percent here, two percent here, two percent here, 19 what it means is that when you look at an actual 20 individual record, there might be 5,000 columns, but 21 there is only two observations, three observations of 22 the information about them, which is not sufficient to Page 62 1 profile those people. 2 And also, when you look at the correlation 3 values of those features against things like Democrat 4 or Republican Party affiliation, the highest 5 correlation that I can recall -- and I can provide you 6 guys with this -- was around .2, which is not reliable, 7 as a value. 8 9 So, the problem that I have with their narrative is that, unless they somehow did something in 10 the couple of weeks in between leaving Ted Cruz's 11 campaign and beginning Donald Trump's campaign that 12 somehow magically fixed this problem, I don't see how 13 they could have done it. 14 15 If that -- does that -- and they haven't actually provided an answer to that. 16 17 18 19 20 MR. HIMES: Okay, thank you. I will yield back. MR. SCHIFF: Ms. Speier, just one last question along the lines Mr. Himes was asking. You know, one of the talking points, one of 21 the alternate narratives is the data we got from 22 Cambridge Analytica wasn't really that valuable. We Page 63 1 actually relied far more on the Republican National 2 Committee for their data. 3 4 Now, the RNC, I don't think, has any kind of data sets like we're describing, or -- 5 MR. WYLIE: 6 MR. SCHIFF: 7 MR. WYLIE: 8 MR. SCHIFF: 9 10 11 12 13 Party does, either. Not to my knowledge. -- you're describing. Yeah. I don't think the Democratic But I wanted to ask you what you make of that argument. Is that just an effort to minimize the significance of the work that was done? MR. WYLIE: So one of the observations that I 14 had whilst I was there -- because I -- during the 15 latter end of my time at Cambridge Analytica they 16 started introducing me to a lot of Republican 17 consultants. 18 me eventually to leave the company. 19 That was one of the things that prompted But one of the things that I observed was the 20 tribalism that existed within the various sort of 21 consulting class of the Republican Party. 22 the things that they didn't like sort of internally was And one of Page 64 1 newcomers, particularly foreign newcomers, coming in 2 and telling them what to do. 3 So with respect to Ted Cruz's campaign and 4 some of the various other campaigns that they worked on 5 early on, there was a lot of tension there because, you 6 know, you have a new company essentially encroaching on 7 their turf. 8 made. That was one of the observations that I 9 The other thing to keep in mind also is that 10 the company doesn't necessarily always do its work for 11 clients, per se. 12 whilst I was there was work that Mercer or Bannon 13 wanted to fund, but didn't necessarily have an overt 14 client. 15 So one of the things that I saw And that, I think, is in part, when you're 16 looking at the broader perspective of what this company 17 ultimately was for, in terms of changing culture, you 18 don't need a candidate to do that. 19 PAC to do that. 20 research and put stuff out onto the Internet, you can 21 go and do that. 22 You don't need a If you just want to go out and do And so, one of the things to just remember is Page 65 1 that it's not necessarily true that all of their work 2 happened officially for the client, and that some of 3 the work would happen sort of in parallel to the client 4 -- also in part because if you do stuff in parallel you 5 don't have to ask permission of the client to go and do 6 it. 7 MR. SCHIFF: Ms. Speier? 8 MS. SPEIER: You have said over and over that 9 10 Cambridge Analytica did this and Cambridge Analytica did that. 11 MR. WYLIE: 12 MS. SPEIER: Yeah. But if I understand correctly, 13 there is no office for Cambridge Analytica in the 14 United States, and no employees. 15 16 MR. WYLIE: They might -- they have an MS. SPEIER: Okay, but there is -- the address. 17 18 functionality of Cambridge Analytica took place in 19 London. 20 MR. WYLIE: Yeah, yeah. The -- so the 21 overwhelming majority -- there were some U.S.-based 22 staff who were American. But they were in a very -- Page 66 1 that was a very small minority of the people. 2 their job was mostly sort of, like, organizing travel 3 and arrangements when people were coming to the United 4 States. 5 positions, with the exception of Steve Bannon, who is 6 American and was also working in a senior capacity. 7 And So they weren't necessarily in senior But the overwhelming majority of staff members 8 were based in London or -- all of them were paid by 9 SCL. And they were at the SCL office, which also 10 converted into the Cambridge Analytica office whenever 11 you wanted to change the letterhead. 12 two sets of letterhead. 13 MS. SPEIER: 14 MR. WYLIE: Like there were Got it. Two sets of email, two sets of 15 letterhead, two sets of everything. 16 recall, there were -- you had different numbers, so one 17 card would say this and the answering machine would be 18 Cambridge Analytica -- 19 20 MS. SPEIER: There were -- as I The vast majority of persons that worked at SCL were foreign nationals. 21 MR. WYLIE: 22 MS. SPEIER: At least 90 percent. At least 90 percent. Page 67 1 MR. WYLIE: 2 MS. SPEIER: At least 90 percent. Were any of them embedded in the 3 Trump campaign, that you know of? 4 after you left, correct? 5 MR. WYLIE: Yeah. Because this was My understanding was that 6 there were data scientists -- and I've seen photographs 7 of this -- who were at the Alamo site, the site in 8 Texas, who were not American citizens, but were working 9 for Cambridge Analytica. 10 11 12 And if -- I can try to get you some of those photos, if -MS. SPEIER: Okay, that would be great. 13 Alexander Nix, after the expose in England, was, I 14 guess, fired or terminated from -- 15 MR. WYLIE: 16 MS. SPEIER: 17 Analytica. 18 still at SCL? 19 Suspended. Suspended from Cambridge But since it's really just a shell, is he MR. WYLIE: That I do not know. What I do 20 know is that he wasn't terminated, he was suspended. 21 And I've been approached by their internal 22 investigator, so I know that they're doing some kind of Page 68 1 internal investigation. 2 But my understanding is that he hasn't 3 actually left the company, he's merely on leave. 4 5 MS. SPEIER: SCL? 6 7 Who is the biggest shareholder of MR. WYLIE: Of SCL? I can't say definitively, but I can actually get you that information. 8 MS. SPEIER: 9 MR. WYLIE: 10 MS. SPEIER: 11 the names of those -- 12 MR. WYLIE: Could you get it? Yeah. I mean if you can get us any of Oh, for -- yes, yes. 13 it's -- we can just get it from -- yeah. 14 fine, yeah. 15 16 MS. SPEIER: MR. WYLIE: 18 MS. SPEIER: The Internet Research Agency had Yeah. And they bought lots of Facebook ads. 20 MR. WYLIE: 21 MS. SPEIER: 22 Yeah, that's -- are you familiar with it? 17 19 In fact, Yeah. They were peculiar, for many of us that looked at them, because they didn't seem to be Page 69 1 necessarily promoting one candidate or another. 2 were trying to create universes of people who were 3 supporters of Black Lives Matters or gun rights 4 supporters. 5 this kind of data mining to then be able to -- 6 They Is that a typical technique that's used in MR. WYLIE: Yeah. So one of the things that's 7 important to remember is that the origin of Cambridge 8 Analytica came from SCL Group, which was -- is a 9 military contractor. And so, when you look at 10 information operations for military projects or 11 counter-extremism projects, you -- there are certain 12 techniques and approaches that you would use that you 13 wouldn't necessarily use normally in any other context. 14 So, for example, if I am the U.S. military, I 15 can't send a marketing campaign to ISIS, because if it 16 says sponsored by the U.S. Army or Uncle Sam says no, 17 right, that's not a credible message. 18 standard marketing or advertising approach to interfere 19 with the operations of ISIS. 20 So I can't use a So, instead, there is something called 21 informational dominance, which is a sort of a -- one of 22 the -- a key concept, which is that in your opponent's Page 70 1 universe you want to gain access to as many channels of 2 information that affect the decision-making of your 3 target as possible so that you can inject information 4 into those channels, and then exploit mistakes or 5 altered perceptions of that target. 6 When you look at, for example, the Internet 7 Research Agency or, more generally, some of the things 8 that Cambridge Analytica was doing, the goal was to 9 change a person's perception of what's happening. 10 So the difference between that approach and a 11 standard political ad is that when you send out an ad 12 from your campaign or from your PAC, it is very overt 13 in the sense that I am trying to convince you of 14 something, this is what my message is. 15 have to even label it as a political act. 16 source of it is apparent. 17 And often you But the The reason that is important is because when 18 somebody looks at it they know that they're seeing an 19 ad. 20 If you want to change someone's perception of 21 something, you will send them information that looks -- 22 that is not branded in that way, and that may look like Page 71 1 a blob, that may look like a news source, that may look 2 like just a chat room with regular people talking. 3 And if you take this notion of gaining 4 informational dominance, and you take your target and 5 you put -- you inject sort of ads, or blogs, or various 6 digital context everywhere where they travel through 7 the Internet, whether you're clicking or -- whether 8 they're on a search engine, whether they're on 9 Facebook, they start to see all of this information, 10 and it starts to change how they think about a 11 particular issue, or what is real and what is not real. 12 And so that is something that is quite 13 problematic with the approach that the company uses, 14 because that is a fundamental denial of someone's 15 agency, because if they start to think that something 16 is real when it's not, then they aren't making a free 17 choice when they're going and voting, because they're 18 basing it on a warped perception. 19 And so, fundamentally, what I have a real 20 problem with ethically about how the company operates 21 is that warping someone's perception may be appropriate 22 if you're trying to interfere with an extremist Page 72 1 organization's recruitment ability, or that -- you 2 know, if you're in a conflict situation. 3 But when you take that same approach and apply 4 it to an election, you know, it -- on one hand, agency 5 doesn't matter because you're -- this is the enemy. 6 the other hand, you're treating every single voter in 7 the same way that you would treat the enemy, and you're 8 warping their perception, which is -- does that -- 9 10 MS. SPEIER: MR. SCHIFF: MR. WYLIE: 14 MR. SCHIFF: MR. WYLIE: 17 MR. SCHIFF: 18 rotating in and out. 19 MR. WYLIE: 20 MS. SPEIER: 22 Okay. We're going to continue to go through the votes, unless you wanted a break. 16 21 One -- I just wanted to let you know we have a series of three votes. 13 15 We've been told that professional -- 11 12 Yes. On Sure. And -- but Members will be Okay, sure, sure. Okay. Russia. MR. WYLIE: Yes. So Professor Kogan in Page 73 1 2 MS. SPEIER: Do you know anything about his work there? 3 MR. WYLIE: 4 MS. SPEIER: 5 MR. WYLIE: Yes. Could you tell us about it? Sure. So he was based -- so, with 6 respect to his Russian work, he was based at St. 7 Petersburg University. 8 online trolling and something called dark triad traits, 9 which is Machiavellianism, psychopathy, and Narcissism. 10 11 12 13 I can go -- explain more about what that is, if you would like, or does that -MS. SPEIER: MR. WYLIE: 15 MS. SPEIER: 17 18 So in St. Petersburg as well, was he at the Internet Research Agency? 14 16 And the research focused on He was -Was he in contract with them at all, or engaged with them at all? MR. WYLIE: The Internet Research Agency? That I don't know. 19 What I do know is that he was flying back and 20 forth between London and Russia at the time that he was 21 managing the data harvesting operation. 22 with various people in Russia about the project that he He was meeting Page 74 1 was doing. 2 I was told at the time that he was giving sort 3 of academic talks, or academic meetings about it, about 4 the processes and the size of the data and what they're 5 doing with it. 6 MS. SPEIER: 7 MR. WYLIE: 8 MS. SPEIER: And how often would he go to MR. WYLIE: When I was there it was several 9 This is what year? This was 2014. Russia? 10 11 times. 12 go back into my email and count the number of times 13 that he said he was in Russia if that's helpful. 14 I couldn't give you an exact number, but I can MS. SPEIER: Okay, that would be helpful. So 15 -- but you have no reason to know that there was any 16 relationship that existed either with SCL or Mr. -- or 17 Professor Kogan and the Internet Research Agency, but 18 they were in the same city. 19 MR. WYLIE: 20 in particular. 21 concerns are -- 22 Not the Internet Research Agency But I can explain more broadly what my MS. SPEIER: Okay, go ahead. Page 75 1 MR. WYLIE: 2 So first, the -- we have -- 3 MS. SPEIER: 4 We've just got to make sure we get in for the vote. 7 8 Figure out how much time we have left in the vote. 5 6 -- with respect to Russia. MR. WYLIE: Oh, no worries. Voting is important. 9 MR. SCHIFF: Just pause one second -- 10 MS. SPEIER: Okay, all right. 11 Let's -- we will go vote and come back, all right? 12 MR. WYLIE: 13 MS. SPEIER: 14 MR. WYLIE: 15 MR. SCHIFF: We will be back in about -- 16 MR. WYLIE: Yeah, no worries. 17 (At 4:16 p.m., a brief recess was taken.) 18 REP. SCHIFF: Ms. Speier? 19 REP. SPEIER: So you were telling us the 20 21 22 Sure. Just hold your thought. Sure. Russia connection. MR. WYLIE: Sure. So just to recap, we had -- we have -- at the time that -- sorry. Page 76 1 2 3 MS. ALLEN: I was just reminding you where you were, my concerns. MR. WYLIE: Oh, right. Sorry. So at the time 4 of the Facebook harvesting project, Dr. Kogan was also 5 working back and forth between London and St. 6 Petersburg -- he was also going to Moscow frequently -- 7 on these Russian research projects. 8 9 So those projects focused on researching online trolling. So trolling is sort of vindictive or 10 nasty sort of bullying-type comments that -- in an 11 online environment, and looking at the underlying 12 psychology of that, and, more generally, behavior on 13 the internet of people, in particular in relation to an 14 area of research called the Dark Triad. 15 Dark Triad are three particular personality 16 traits -- Narcissism, Machiavellianism, and 17 Psychopathy. 18 trolling and the Dark Triad traits and profiling 19 literally the worst traits in people. 20 So his research in Russia was on online So SCL Group was aware of this research and 21 went around and spoke about it to different clients 22 that they had around the world. Page 77 1 So I can provide you with, for example, one 2 email to the Minister of National Security of Trinidad 3 talks about the quote “very interesting work” that 4 Aleksandr Kogan is doing for the Russians in relation 5 to profiling of particular targets or the Trinidadian 6 government. 7 As that was -- as -- so as Dr. Kogan was going 8 back and forth to Russia, at the same time we got 9 approached -- and I don’t know -- unfortunately, I 10 don’t know why we got approached, but we got approached 11 by a company called LUKOIL. 12 largest oil company in Russia. 13 So LUKOIL is the second- We got approached by fairly senior executives 14 at the company. 15 presentation to them, which I can -- I can pass on to 16 you -- the British committee already has it -- where he 17 talks about rumor campaigns, voter inoculation or 18 attitudinal inoculation, so changing people’s minds in 19 a way that they don’t quite understand that they’re 20 being manipulated, for lack of a better word, and, more 21 generally, undermining confidence in civic 22 institutions, with reference to projects that they had And Aleksandr put together a Page 78 1 worked on in Africa. 2 Alexander Nix said that the LUKOIL pitch was 3 for a consumer loyalty project in Turkey, but to me I 4 have, you know, no recollection of that. 5 don’t know why he would give a presentation about rumor 6 campaigns and undermining faith in civic institutions 7 in Africa for a consumer loyalty project in Turkey. 8 And also, I One of the other things that I can provide to 9 you is emails that reference sending white papers that 10 I had written on the internal capacity of our American 11 data assets and the algorithms and profiling that we 12 were doing and sending that to senior executives at 13 LUKOIL, including a reference to the CEO of LUKOIL. 14 One of the concerns that I have is that LUKOIL 15 has fairly well-known relationships with the FSB, has 16 an information-sharing agreement with the FSB. 17 that in Europe it often operates in places that are 18 difficult to get FSB agents in or going around 19 collecting information under the auspices of being a 20 company. 21 22 I know The other thing that I should mention is that on the -- on the defense side of the company, SCL does Page 79 1 a lot of work in Eastern Europe, in particular in the 2 Baltics. 3 often involve counter-propaganda efforts. 4 counter-propaganda for NATO clients. 5 That research and -- or, rather, operations So Russian And so my other concern is that given that the 6 company was working in Eastern Europe on the military 7 side for NATO clients that we probably would have been 8 some kind of intelligence target, given that we were 9 already operating on their turf, and that Alexander Nix 10 told the company, in no uncertain terms, about the 11 assets that we were building in the United States, what 12 our capacity was, how it could be used, makes reference 13 to rumor campaigns and, you know, disinformation and 14 undermining civic -- confidence in specific 15 institutions, and then informs the company that, in 16 fact, the -- the leader manager of that data harvesting 17 is actually going to Russia. 18 And that -- the concern that I have is not 19 that Cambridge Analytica somehow cut a deal with the 20 Russians and there is a grand conspiracy. 21 that I have is that they made a lot of noise for a long 22 time. The concern It is -- you know, from 2014 to -- you know, Page 80 1 essentially from the spring of 2014 to the spring of 2 2015, so almost a year, with LUKOIL, and that if -- 3 REP. SPEIER: 4 MR. WYLIE: “They” being? Cambridge Analytica. And that if 5 that information was passed on to any number of 6 agencies in Russia, given that the data that was being 7 acquired through the Facebook harvesting scheme was 8 often not encrypted, that even if it was encrypted 9 could have been acquired through something as simple as 10 a key logger. 11 So key logger is either a device or a piece of 12 software that is uploaded to a computer that then logs 13 what you’re typing, which then you can use to get, you 14 know, username, password, all of that. 15 And that given that Dr. Kogan was going back 16 and forth to Russia, as we’re making it known to -- 17 rather, as Cambridge Analytica is making it known to 18 LUKOIL that these projects are going on, the concern is 19 that it would have been extremely easy to get access to 20 this data. 21 on Dr. Kogan’s computer in Russia. 22 Extremely easy. As simple as a key logger As easy as sending -- the data science teams Page 81 1 at SCL, when I was there, the majority of them were 2 Eastern European, and they did not do background checks 3 on any of them. 4 European you are somehow, you know, in -- working for 5 -- but that the security practices at the time were 6 incredibly poor. 7 Not to say that if you’re Eastern I can give you examples of where the company 8 would be emailing data sets on hundreds of thousands of 9 American citizens in unencrypted emails, which would 10 have been incredibly easy to access. 11 The other sort of -- 12 REP. SPEIER: And that would be to a contract 13 that they had with a company where, in the United 14 States or somewhere else? 15 MR. WYLIE: 16 REP. SPEIER: 17 they would be sending all this -- 18 MR. WYLIE: With -- a contract with -A company they had -- you said So the Kogan project, so, for 19 example, between Aleksandr Kogan or some of his 20 researchers and Cambridge Analytica. 21 “Hey, can I get” -- 22 REP. SPEIER: Got it. So if we said, Okay. Page 82 1 MR. WYLIE: -- “a data set,” rather than 2 sending it through a secure FTP site, for example, it 3 would just be emailed as just an open, unencrypted 4 file. 5 REP. SPEIER: 6 with SCL or CA or -- 7 MR. WYLIE: Professor Kogan was a contractor He operated through a company 8 called Global Science Research. 9 individual, wasn’t contracted. So he, as an It was his company, 10 GSR. GSR was made up of Dr. Kogan and then another 11 gentleman, Joseph Chancellor, who now works at 12 Facebook. 13 REP. SPEIER: 14 MR. WYLIE: What’s the name? Joseph Chancellor. I can send you 15 -- I can send you the name. 16 that happening, one of the other things that I just 17 find slightly bizarre about what Cambridge Analytica -- 18 some of the research streams that got set up after 19 Bannon took over, the only tested foreign leader in the 20 research that was being done in the United States was 21 on Vladimir Putin. 22 So in addition to all of And also, the -- one of the primary focuses of Page 83 1 the research in relation to Americans’ perspectives on 2 foreign issues was Russian expansion into Eastern 3 Europe and whether Vladimir Putin was justified in 4 taking over parts of Ukraine. 5 REP. SPEIER: 6 MR. WYLIE: Okay. And so I can provide you with, for 7 example, transcripts of focus groups on that. 8 the thing -- the thing that I just find -- 9 REP. SPEIER: 10 MR. WYLIE: 11 REP. SPEIER: 12 And so Hold that thought. Yeah. We’re down to 50 votes. We’ll go off the record just for a minute. 13 (Recess taken from 4:34 p.m. until 4:47 p.m.) 14 REP. SCHIFF: Ms. Speier. 15 REP. SPEIER: Where did we leave off? 16 MR. WYLIE: Oh, right. 17 I can’t read your writing. 18 Sorry. 19 I was talking about -- I’m sorry. Oh, right. So the -- so I’ll just back up a little bit. 20 So one of the -- the only foreign leader that was 21 tested -- 22 REP. SPEIER: Oh, right. Page 84 1 2 MR. WYLIE: doing was Vladimir Putin. 3 REP. SPEIER: 4 MR. WYLIE: 5 it. 6 speculate. 9 10 11 12 And who was requesting that? I -- I don’t know who requested I know that I didn’t, and so I would rather not 7 8 -- in the research that we were REP. SPEIER: But it was also Vladimir Putin and -MR. WYLIE: REP. SPEIER: And --- people’s feelings about the annexation of Crimea? MR. WYLIE: Yeah. So -- so we -- being 13 Cambridge Analytica, in focus groups and in message 14 testing -- and I believe that one of the models that 15 Dr. Kogan provided to Cambridge Analytica actually 16 modeled views on Russia, in terms of foreign relations, 17 because that was one of the message testing -- 18 REP. SPEIER: You just said focus groups. 19 Were you doing focus groups -- 20 MR. WYLIE: 21 REP. SPEIER: 22 MR. WYLIE: There were --- in the U.S.? Yes. There was both a qualitative Page 85 1 and quantitative stream of research happening in 2 parallel. 3 4 So -REP. SPEIER: All of that information is now housed within SCL, right? 5 MR. WYLIE: Yes. I can provide you some of 6 the transcripts because I still -- I still have some of 7 the transcripts from the -- from the focus groups, for 8 example, if you’d like that. 9 The focus groups were used -- so qualitative 10 research is not usually generalizable, so -- because 11 your sample size is, you know, 12 people in a room 12 talking. 13 affect what -- the output of that. 14 So there’s all kinds of things that could But one of the reasons that focus groups and 15 in-depth interviews were used was essentially to come 16 up with better -- essentially asking people, what is it 17 that I don’t know about you that I should know about 18 you? 19 from that to then test quantitatively in a way that you 20 can actually validate through control groups and -- And then coming up with theories or hypotheses 21 22 REP. SPEIER: Bannon -- In your conversations with Mr. Page 86 1 MR. WYLIE: Yeah. 2 REP. SPEIER: -- did he indicate to you any 3 interest in wanting to suppress the vote in the United 4 States on some level? 5 MR. WYLIE: So if by “suppression” you mean 6 discourage particular groups of people or discouraging 7 the likelihood of particular groups of people from 8 voting, is that what you mean by “suppress”? 9 10 REP. SPEIER: MR. WYLIE: Yes. Then yes. So one of the things 11 that I could provide to you is a memo that references 12 voter disengagements in bold as an objective for some 13 of the American operations. 14 My understanding of that was that that quote 15 “voter disengagement” would be focused on particular 16 groups of people who are more prone to vote Democrat. 17 So it wasn’t -- it wasn’t general. 18 -- to that. 19 It was specific to So yes. And with respect to your question about Steve 20 Bannon in particular, that was something that came up 21 in conference calls with him. 22 REP. SPEIER: How about efforts to influence Page 87 1 2 voters on race? MR. WYLIE: So one of the things that emerged 3 from some of the research was that there was sort of an 4 undercurrent of certain -- certain types of Americans, 5 typically white men who -- but there were also groups 6 of women, though -- who -- are you familiar with the 7 concept of race realism? 8 9 No. So race realism is a movement within the alt right, which is the idea that you can call something -- 10 you can call someone racist for having prejudicial 11 views towards particular ethnicities or races. 12 But a defense to that -- and so just to be 13 clear, I’m not personally defending it, I’m just -- to 14 explain it -- would be that it’s simply being quote 15 “realistic about the realities of race,” and that there 16 are certain natural hierarchies to people. 17 And this is something that I know Steve Bannon 18 was -- found interesting, as something that he wanted 19 to explore. 20 REP. SPEIER: So at the time that you were 21 engaged with him, he didn’t have a candidate in the 22 U.S. election, or did he? Page 88 1 MR. WYLIE: There were -- there were lots of 2 candidates that Robert Mercer or Rebekah Mercer wanted 3 to help or support. 4 irrespective of the candidate, because the first step 5 to creating the company’s assets would be to create 6 insights, models, targeting, message testing, that 7 would be applicable anywhere in the United States for 8 whichever candidates, you know, the Mercers wanted. 9 But the underlying research was So, but to answer your question, there were 10 candidates that Robert Mercer wanted from the very 11 beginning to support, or various PACs or various sort 12 of movements and things like that. 13 14 REP. SPEIER: articulated to you? 15 16 MR. WYLIE: REP. SPEIER: 18 MR. WYLIE: 20 Yeah. I can provide you with a list of -- 17 19 And did he -- were those Okay. That would be great. -- the -- and, I mean, some of them are currently still in Congress, for example. REP. SPEIER: 21 you want us to know? 22 MR. WYLIE: What is the most important thing I want you to -- so what’s -- Page 89 1 what’s really important for me to speak with you as 2 representatives of the United States is that a lot of 3 -- a lot of the wrongdoing that this company has done 4 was in the United States. 5 So, you know, the bulk of the misappropriated 6 Facebook data was on American citizens. The bulk of 7 the research was on American citizens. The -- I want 8 to inform you collectively about -- that this company, 9 and more broadly the backers of this company, so 10 whether it’s, you know, the Mercers or Steve Bannon, 11 are willing to manipulate and work the perceptions of 12 American voters, simply to exploit -- exploit that 13 warped view in an election. 14 And the problem that I have with that is that 15 that is fundamentally incongruent with a free election. 16 If you are trying to trick people, for lack -- to 17 really simplify it, trick people into voting for you, 18 by changing what they think is real, that is wrong, and 19 I would like people in the United States to know that. 20 I also think that there may be grounds to 21 investigate the company, not just in the UK but also in 22 the United States, with respect to electoral law Page 90 1 infractions and potentially Foreign Corrupt Practices 2 Act and other sort of laws specific to the United 3 States, which I’d like to inform you about. 4 And then also what is particularly concerning 5 to me is the interactions that I witnessed between the 6 company and LUKOIL, and also the work that Dr. Kogan 7 was doing in Russia, and that given that this company 8 was harvesting a massive amount of very sensitive data 9 on a lot -- you know, tens of millions of Americans, 10 and they were doing it whilst they were telling very 11 senior executives of the second-largest oil company in 12 Russia about it, as the -- the lead researcher who is 13 in charge of that harvesting was going back and forth 14 to Russia. 15 You know, as some of his colleagues -- one of 16 the things that I can provide to you is some of his 17 colleagues were also going to Russia, talking to 18 politicians in Russia about some of the work that the 19 psychometric center was doing. 20 concerned that something -- that this hasn’t been 21 investigated, and that it really should be. 22 So I am just generally And given that, you know, this is -- that this Page 91 1 is a company that also, you know, has a long history of 2 working with the American military, from an American 3 perspective, you know, is this an appropriate company 4 that your government should be engaging on those kinds 5 of projects? So if that answers your question. 6 REP. SPEIER: Thank you. 7 REP. SCHIFF: We’d be very interested in any 8 documentation you have on the presentations to LUKOIL, 9 on the -- 10 MR. WYLIE: 11 REP. SCHIFF: 12 13 Yeah. -- other Kogan employees who were -- also interacted with Russia. MR. WYLIE: So, for example, one of -- one of 14 his colleagues -- and I have the screen cap of it -- 15 posted that he was going to meet senior politicians in 16 Russia to talk about the research that they were doing, 17 and he mentions the Prime Minister of Russia as 18 somebody that he was going to give a talk to. 19 REP. SCHIFF: If I could, and then I’ll turn 20 to Mr. Swalwell. Have you seen any connections between 21 the work that Cambridge Analytica was doing to analyze 22 American attitudes towards Putin and to the annexation Page 92 1 of Ukraine -- 2 MR. WYLIE: Yeah. 3 REP. SCHIFF: -- and the Trump campaign and 4 LUKOIL? 5 between the Trump campaign, that particular -- the 6 particular interest in that research as well as the 7 interactions with LUKOIL with Russia generally? 8 9 In other words, anything that crosscuts MR. WYLIE: Not with LUKOIL in particular. More broadly, when you look at a lot of the narratives 10 that the Trump campaign was -- was putting out. 11 were a lot of the messages and general sort of concepts 12 or imagery that were being tested at Cambridge 13 Analytica. 14 15 16 These So -- REP. SCHIFF: And were those being tested in 2014? MR. WYLIE: Yeah. So, for example, drain the 17 swamp, images of walls, you know, the -- you know, my 18 -- one of the -- the wall is really interesting because 19 I don’t actually think that their intention is really 20 to build a wall, actually. 21 I think that part of -- so there is -- in 22 post-reunification Germany, there was some research Page 93 1 done on something called -- and my German isn’t -- I 2 don’t speak German, but something to the effect of 3 Mauer im Kopf, which means the wall in the head. 4 So Germans overestimate the distance east to 5 west compared to north to south of the country. 6 they imagine their country more stretched out than it 7 actually is. 8 imagery of walls, which was, you know, essentially 9 actually only in Berlin. 10 11 So And part of that had to do with the It wasn’t the entire country that had a giant concrete wall. But that -- that image created a psychological 12 distancing between the two sides of the country. 13 that even after that physical wall came down, that 14 mental wall remained and is enduring even to this day. 15 People still imagine the country as wider than it 16 actually is. 17 And And that if you are Steve Bannon, and you are 18 looking to close off the United States so that it can 19 quote/unquote “find its purpose again,” it is -- that’s 20 a direct quote. 21 creating this psychological distance thing from, you 22 know, the rest of the world is really important. So if that is one of your objectives, Page 94 1 And so this notion of building a wall isn’t 2 necessarily just to actually build a physical wall. 3 It’s -- and I don’t think it has anything to do with 4 immigration. 5 the rest of the world and creating a very strong, 6 enduring symbol of that. It’s to do with distancing Americans from 7 Sorry, that was a bit of a tangent, but -- 8 REP. SCHIFF: 9 No, no. That’s very interesting. 10 Mr. Swalwell. 11 REP. SWALWELL: 12 And, Mr. Wylie, thank you for coming in with Thank you. 13 us. 14 working with Cambridge Analytica. 15 Can we I guess go back to your last day at SCL or MR. WYLIE: Why did you leave? After Steve Bannon took over, the 16 sort of internal company culture became quite toxic. 17 When I first met Steve Bannon, I actually really liked 18 him. 19 guy to talk to, and you can have conversations like, 20 what is culture, or talk about intersectional feminism 21 and fractured identities in the nature of self, and he 22 will fully engage you in that conversation. He is very interesting -- he’s a very interesting So he’s a Page 95 1 very interesting guy to actually talk to. 2 But when you become his subordinate, you are 3 then treated as such. 4 you know, it was -- became very difficult to work in. 5 And then, also, being instructed to -- to apply some of 6 the research that we were doing originally to the 7 United States not only just applied to the United 8 States, but then being introduced to, you know, his 9 associates, who had, in my view, extremely warped 10 political views that I was genuinely uncomfortable 11 with. 12 So the culture of the company, Sitting, for example, in a meeting with my 13 colleagues, and essentially talking about, you know, to 14 a group that we were introduced to, about essentially 15 how to exacerbate or reactivate homophobia in the 16 United States when many of my colleagues were gay 17 themselves, and it just became clear that this is not 18 something that I can do. 19 I don’t -- you know, and I never ended up 20 signing the final contracts and intellectual property 21 agreement that the company wanted me to sign, which 22 would have locked me in to the company, because I just Page 96 1 couldn’t imagine spending 10 years doing that. 2 couldn’t imagine like my career doing that. 3 4 REP. SWALWELL: What was the total amount of time you worked there? 5 MR. WYLIE: 6 REP. SWALWELL: 7 give them a letter? 8 that play out? 9 I MR. WYLIE: About a year and a half. How did you leave? Did you tell someone? Did you How did I sent an email to Alexander Nix, 10 and I -- I didn’t fully go into the reasons why I 11 wanted to leave. I just said, in July 2014, that I’m 12 going to leave. But the agreement that I had was that 13 I would phase out my work until November, which is the 14 -- when the midterm was happening; and then after that 15 point, no more. 16 REP. SWALWELL: 17 MR. WYLIE: Did he respond to you? He was -- in person he did, and he 18 took me to -- to lunch and tried to convince -- so he 19 just -- for Alexander Nix and for Steve Bannon, I think 20 they found it really difficult to understand -- like if 21 we just pay you more, why won’t you stay? 22 REP. SWALWELL: Why won’t -- Did they offer to pay you Page 97 1 more? 2 MR. WYLIE: 3 REP. SWALWELL: 4 MR. WYLIE: Yeah. Okay. I got offered to pay -- be paid a 5 lot more. 6 I can’t imagine myself doing this anymore.” 7 not -- when I -- when I signed up originally, we 8 weren’t doing anything with the alt right. 9 And, you know, I just -- I said, “Literally, This is We weren’t -- I didn’t even know who Steve 10 Bannon was, and, you know, the thing that was appealing 11 to me is that I could work on projects that, you know, 12 if it related to counter-extremism, or behavior change 13 in Africa with respect to HIV-prone behaviors, or 14 things like that, to me that was morally acceptable. 15 What they were doing I didn’t find acceptable, 16 and I really didn’t like that I was playing a 17 foundational role in it. 18 that I either leave now or I’m going to be locked in, 19 and I just said, “Don’t want to” -- 20 21 22 REP. SWALWELL: And I essentially understood Were you ever under a nondisclosure agreement with -MR. WYLIE: Yes. I -- Page 98 1 REP. SWALWELL: 2 MR. WYLIE: Which company? I have a nondisclosure agreement, 3 a mutual nondisclosure agreement with SCL, and then 4 after I left they then threatened various legal actions 5 and made all kinds of accusations, which were actually 6 just completely untrue, and chased me for months. 7 REP. SWALWELL: 8 nondisclosure agreement with SCL? 9 MR. WYLIE: When did you sign the I signed the nondisclosure 10 agreement with SCL in 2013, and then after receiving 11 legal threats in 2015 after I left, I then signed 12 something called an undertaking of confidentiality, 13 which I believe is signed as a deed. 14 it’s a slightly higher tier of duty of confidence, as 15 it were. 16 REP. SWALWELL: 17 MR. WYLIE: So there is -- Why did you sign that? Well, they just wouldn’t stop 18 sending letters to my lawyers, and then they threatened 19 to take me to court over -- 20 REP. SWALWELL: 21 sending the letters? 22 MR. WYLIE: Who is “they”? SCL. Who was Page 99 1 2 REP. SWALWELL: employees of SCL personally contacting you? 3 4 MR. WYLIE: Oh, lots. REP. SWALWELL: 6 MR. WYLIE: 7 agreement, no. 8 company. 10 About -- not signing the Just about what was happening in the REP. SWALWELL: Okay. Who was pressuring you at SCL to sign the agreement? MR. WYLIE: 12 REP. SWALWELL: 14 About signing an agreement? Oh, no. 11 13 Yeah, everybody was talking to me. 5 9 And was it -- were any Alexander Nix and Steve Bannon. And were they doing that in person, on the phone, electronically? MR. WYLIE: All of the above. And it was 15 because -- the problem was that Robert Mercer put in 16 money, and Alexander Nix sort of sold intellectual 17 property to -- or, rather, SCL sold intellectual 18 property to Cambridge Analytica, but they didn’t fully 19 finalize or execute agreements with myself about the 20 intellectual property that I was generating. 21 22 The mutual nondisclosure agreement that I have with the company very clearly states that any Page 100 1 intellectual property that I develop I retain, unless I 2 specifically license it to the company. 3 found out, he obviously was concerned about that, 4 because their primary concern at the time was that they 5 thought that I might go and take this and go and like 6 work on the Hillary campaign or something and try to 7 replicate. 8 This is -- this is -- 9 REP. SWALWELL: 10 MR. WYLIE: When Bannon Did they tell you that? It came out in conversations 11 between my lawyers and their lawyers that that was one 12 of the concerns that they had. 13 part, why they have this narrative that I was a 14 disgruntled employee and that I was trying to set up a 15 rival firm. But -- and this is, in 16 I never set up a rival firm. 17 party in this entire, you know, saga that never 18 commercialized the data. 19 didn’t set up a rival firm. 20 21 22 REP. SWALWELL: I’m the only I never sold the data. Can we go back to the undertaking? MR. WYLIE: Yeah. Sure. Sorry. I Page 101 1 REP. SWALWELL: Is there anything about that 2 undertaking deed that you signed, or the mutual 3 nondisclosure agreement that you signed -- 4 MR. WYLIE: 5 REP. SWALWELL: 6 7 Yeah. -- that is limiting your testimony to us today? MR. WYLIE: No. I am -- my entire 8 conversation with you is breaking in. But in the UK at 9 least, there is a defense to breach of confidence, 10 which is if you are revealing potential criminality or 11 grossly unethical behavior, you have a defense to 12 breach of confidence. 13 14 15 REP. SWALWELL: Have you been interviewed by the Special Counsel here in the United States? MR. WYLIE: No, I haven’t. Next week I am 16 being interviewed by the FBI, Department of Justice, 17 and the New York Attorney General. 18 REP. SWALWELL: 19 Is that with respect to the Russia investigation? 20 MR. WYLIE: 21 questions they want to ask me. 22 It is with respect to what REP. SWALWELL: I’m sorry. Counsel? Page 102 1 MR. WYLIE: Yeah. They haven’t -- we haven’t 2 -- they don’t give you the questions in advance when 3 they interview you. 4 So it -- REP. SWALWELL: Do you know if it’s -- the FBI 5 interview is from the Special Counsel team, or is that 6 separate? 7 MR. WYLIE: It is being coordinated by the 8 National Crime Agency in the UK. 9 various crimes, or at least from my perspective what So I have reported 10 are possible crimes to investigate, to the NCA, which 11 is the equivalent of the FBI in the UK. 12 The NCA has special relationships with the 13 FBI. 14 the FBI, given what I have handed over to them, to then 15 interview me. They have invited who they as appropriate from 16 REP. SWALWELL: 17 MR. WYLIE: Where will that interview be? It will happen, I believe, on a 18 military base in the UK. 19 REP. SWALWELL: And did you say that it would 20 be a joint interview, or is it separate interviews, New 21 York Attorney General, and then FBI, DOJ? 22 MR. WYLIE: My understanding -- and correct me Page 103 1 if I’m wrong, Tamsin -- is that each respective agency 2 or authority will have an opportunity to ask me 3 questions. 4 the interview also. 5 But the other authorities will be watching MS. ALLEN: Is that -We -- I don’t know -- I don’t 6 think we can give a complete answer to that yet, though 7 we’ve asked questions about how the interviews are to 8 be structured and we don’t yet know completely. 9 REP. SWALWELL: What did you do once you left 10 Cambridge Analytica? 11 like you’ve gone to school. 12 MR. WYLIE: What you done since? Yeah. It sounds I’ve done -- so I know this 13 is going to sound ridiculous. 14 fashion, I have done projects in politics, working as a 15 consultant mostly in data science. 16 projects, for example, for the Canadian government 17 after Justin Trudeau got elected. 18 REP. SWALWELL: So I’ve done projects in So I’ve done Have you done any work in -- 19 that was with a rival of Cambridge Analytica or a rival 20 campaign of a campaign that they were working on? 21 22 MR. WYLIE: No. I have not -- I am -- so this is -- I find this entire narrative very frustrating Page 104 1 because after I -- the reason I left Cambridge 2 Analytica is because I did not want to continue working 3 on Steve Bannon’s vision for the alt right. 4 I, therefore, cannot be a rival to this 5 company that works for Republicans because I have -- 6 and I will be super clear -- I have never worked for a 7 Republican since. 8 conservative candidate since, and I will not. 9 have not -- I am not setting up a rival -- I don’t know I have never worked for a And so I 10 where this rival firm is that apparently I have set up 11 and managed. 12 REP. SWALWELL: Can you tell me, how was it 13 that you stayed in contact with employees at Cambridge 14 Analytica? 15 good window into what was going on all the way up to 16 the Trump data operation, Project Alibi. 17 Because it sounds like you had a pretty MR. WYLIE: So it’s because I -- the people 18 that I brought on when I was at SCL, and then later 19 when the CA was expanding, were people who I had worked 20 extensively on projects in the past in other contexts. 21 22 Most of the people who got brought on early on were not conservatives. You know, we had -- you know, Page 105 1 not to be flippant about it, but we had a lot of vegans 2 working on the project. 3 work on the project. 4 weren’t -- it wasn’t -- it wasn’t a conservative team 5 that got set up. 6 We had a lot of gap people You know, we -- you know, there And when I left, people would continue talking 7 to me, in part because I had a just really positive -- 8 you know, I was friends with some of them -- positive 9 relationship with them, and then also they started 10 getting more and more frustrated with what was 11 happening. 12 13 I left. -- after I left. 14 15 A lot of people started leaving after REP. SWALWELL: on your life? 16 MR. WYLIE: 17 REP. SWALWELL: 18 Not on my life, no. MR. WYLIE: 20 REP. SWALWELL: 22 What kind of -- any physical threats? 19 21 Have you had any threats made Not -- not physical. How about like financial ruin or personal ruin threats? MR. WYLIE: I’ve had -- I think it would be Page 106 1 more appropriate if we spoke about that off record. 2 that possible? 3 I’ve -- REP. SWALWELL: Actually, I just want to make 4 sure, Mr. Schiff, if that’s all right? 5 threats made, and he doesn’t feel comfortable stating 6 them on the record. 7 REP. SCHIFF: 8 REP. SWALWELL: 9 Is He has had That’s fine. Okay. If the stenographer could -- we’ll go off the record momentarily. 10 (Recess taken from 5:15 p.m. until 5:17 p.m.) 11 REP. SWALWELL: 12 Were there any Russians working at Cambridge Analytica, or Russian nationals? 13 MR. WYLIE: My understanding is Dr. Kogan was, 14 if not Russian -- I believe he was either Russian or 15 Maldovan. 16 naturalized as an American citizen. 17 His family is Russian-speaking, and he then REP. SWALWELL: While he was working there, 18 did you observe him traveling back and forth from 19 Russia or any -- 20 MR. WYLIE: 21 REP. SWALWELL: 22 countries? You did. Yes. -- of the Soviet bloc Page 107 1 MR. WYLIE: 2 REP. SWALWELL: 3 MR. WYLIE: 4 Yeah. Oh, yeah. And you did do that. Oh, yeah, yeah. I have email records of that. 5 REP. SWALWELL: 6 MR. WYLIE: Anyone else? There -- I mean, most of the data 7 science team were Eastern European. A lot of them 8 spoke Russian, so it wasn’t unusual to -- to hear 9 Russian in the office. 10 At the time, though, that’s not -- you know, 11 for me, it’s not suspicious that somebody comes from a 12 certain part of the world, or if their colleague also 13 speaks that language, they will speak to them in that 14 language. 15 16 REP. SWALWELL: or the focus groups -- 17 MR. WYLIE: 18 REP. SWALWELL: 19 Yeah. -- about how Putin was viewed -- 20 MR. WYLIE: 21 REP. SWALWELL: 22 Can we go back to the polling commissioned that? Yeah. -- in the United States? Who Page 108 1 MR. WYLIE: 2 REP. SWALWELL: 3 Cambridge Analytica did. Cambridge Analytica to do that? 4 MR. WYLIE: 5 REP. SWALWELL: 6 No. No. We -- Do you know why they wanted to do that? 7 MR. WYLIE: 8 REP. SWALWELL: 9 Was there a client that asked Honestly, no. Is it possible I guess, just based on how you understood Cambridge Analytica to 10 work, that Mr. Nix or Mr. Kogan or somebody who had a 11 relationship with Russia could have been asked by 12 someone in Russia to do that type of research; is that 13 something that would not have been out of the ordinary? 14 MR. WYLIE: Yeah. Yes. My intuition, and so 15 I’m just going to be super clear, this is not me 16 stating a fact. 17 that Dr. Kogan added some of these research areas, 18 because he was quite involved in the actual testing of 19 messages and figuring out ways to test those messages, 20 to test those constructs. 21 22 I’m just stating a suspicion -- is And, you know, his -- it was him that actually went out and organized the actual surveys, testing, et Page 109 1 cetera. 2 in deciding what kinds of constructs we would look at. 3 So he had -- he played a very influential role But I can’t tell you for sure who did it. 4 I can tell you that having gone and asked 5 several former staff members, none of them -- none of 6 the people that I have spoken to commissioned that, or 7 at least told me that they commissioned that area -- 8 9 REP. SWALWELL: guess. Let me ask it this way I Was this something that was out of the 10 ordinary, meaning that your firm was doing, at least 11 ostensibly to you and other employees, research in the 12 United States about how Americans viewed Russia, and 13 there was no known client attached to it? 14 MR. WYLIE: There was that -- that -- as I 15 understand it, at least when I was there, that stream 16 of research was never actually applied to a specific 17 client. 18 It was -- it was just research that was done. REP. SWALWELL: And were there other 19 instances, though, like that where research would be 20 done that didn’t have any known application? 21 that sort of an outlier? 22 MR. WYLIE: Or is Not that -- it’s an outlier. Not Page 110 1 that I can remember because, you know, if you’re -- so 2 I don’t know how it would be useful to a campaign, an 3 American campaign, to know about people’s views on 4 Vladimir Putin. It’s not like you’re running against 5 Vladimir Putin. It’s not like you’re -- it’s not like 6 a major topical issue is Russian expansionism in 7 Ukraine. 8 9 So whereas if you are testing other, you know -- your perspective on feminism, for example, that 10 could be used in all kinds of different -- you know, 11 your -- your perspective on gun rights, your 12 perspective on different things, that can be applied. 13 I do not remember any -- at least when I was 14 there, any example of the Russian research being 15 deployed in any way. 16 REP. SCHIFF: On that point, that focus 17 grouping and qualitative and quantitative analysis on 18 the Putin-Ukraine issue -- 19 MR. WYLIE: Yeah. 20 REP. SCHIFF: -- is that something requested 21 by Bannon specifically, or do you know how that got 22 into the mix? Page 111 1 MR. WYLIE: This is something that I have been 2 trying to figure out, because I know that no one on the 3 message testing team commissioned that area of 4 research. 5 elimination, at least -- and I -- I feel like they 6 would be honest to me. 7 think that they would be blind to me. 8 9 10 So I can -- through a process of There is sort of two people left. either got Steve or you’ve got Kogan. REP. SCHIFF: 12 MR. WYLIE: 13 REP. SCHIFF: MR. WYLIE: 16 REP. SCHIFF: 18 19 20 But I can’t -- I Because this was 2014, right? Yeah. So this is well before the Trump campaign. 15 17 You’ve can’t tell you for sure who would have -- 11 14 So I don’t have any reason to Yeah. And if this is done by either Bannon or Kogan or Nix -MR. WYLIE: I don’t think it would have -- well, I don’t think it would have been Nix. REP. SCHIFF: Well, the reason I ask is, for 21 the very reason you mention, because it’s well before 22 the Trump campaign, yeah -- Page 112 1 MR. WYLIE: Yeah. 2 REP. SCHIFF: -- and Bannon would not have 3 known necessarily that there would be a U.S. 4 Presidential campaign advocating a close relationship 5 with Russia, or mitigating the negative views on Putin 6 and Ukraine, it would make much more sense if it was in 7 connection with a Russian-interested client. 8 9 MR. WYLIE: Yeah. Something just to say to that is that Steve Bannon has an admiration for sort of 10 strong man-type leadership. 11 when I was there and he was talking about sort of the 12 kind of leader that he likes, sort of at the time, in 13 my head, I imagined Mussolini, but I could have also 14 very easily imagined Putin. 15 at the time. 16 of the Steve Bannon archetype of a very loud but strong 17 kind of leader that he liked. 18 So I know that, at least Just it wasn’t in my head But I kind of imagined Mussolini as sort But if I were to -- if I were to -- if I were 19 to -- if I had to bet on who it was, I suspect that 20 because Dr. Kogan was more involved directly in the 21 actual management and commissioning of research, and 22 the actual execution of developing surveys and putting Page 113 1 them out and collecting that data, that he may have had 2 a role in that. 3 4 5 REP. SCHIFF: And that might have been of interest to the people he was talking to in Russia? MR. WYLIE: It could -- I mean, it could be. 6 This is -- this is, again, partly why -- I mean, I 7 don’t have every answer, but partly -- one of the 8 reasons why, you know, I want to speak to you and 9 others is to kind of put that question to the 10 11 authorities and look at finding an answer to that. REP. SCHIFF: And you’ve probably seen the 12 indictment that Mueller returned on the 13 Russians 13 that lays out in detail what the Internet Research 14 Agency was doing in terms of their use of social media. 15 From what you have -- what you see in the 16 indictment, and from what you’ve seen more generally in 17 the open source -- 18 MR. WYLIE: Yeah. 19 REP. SCHIFF: -- how the Russians deployed, 20 advertising organic content, paid content, does 21 anything about it strike you as suggestive of the use 22 of the modeling that Cambridge Analytica did? Page 114 1 MR. WYLIE: The thing that -- and so I’m not 2 an -- like an expert on the -- you know, what the 3 Internet Research Agency was doing, but just generally 4 from what I’ve seen, the kinds of -- when you look at 5 sort of the style of message or sort of just -- even 6 the general sort of aesthetic of it, how it’s slightly 7 bizarre, it feels slightly bizarre, but it’s sort of 8 leading -- it’s quite divisive. 9 -- it felt familiar. 10 And that, to me, felt But I don’t know if the IRA in particular used 11 -- used CA modeling or not, so I can’t answer that as a 12 definitive. 13 have an agency, whether it was the IRA or some other 14 agency that had access to this data, it would have been 15 extremely helpful in either developing narratives and 16 understanding what kinds of narratives could become 17 divisive, if that was your goal, or, secondly, where or 18 how to target it. But what I can say is that if -- if you 19 So I can’t -- I can’t answer that one way or 20 the other, but one of the concerns that I have is that 21 that’s a possibility. 22 somebody could go and take that and look into it in And I -- I would hope that Page 115 1 some way. 2 REP. SCHIFF: 3 REP. SWALWELL: Mr. Swalwell. You talked to members from 4 Cambridge Analytica throughout, you said, the Trump 5 campaign, and you mentioned earlier that foreign 6 nationals had gone to San Antonio. 7 kind of foreign nationals they were? 8 9 10 MR. WYLIE: Do you know what I don’t want to -- I don’t want to misspeak, so I -- I believe that they were -- what I -how about if I go back and -- 11 REP. SWALWELL: 12 MR. WYLIE: Sure. -- I get the photos that I’m 13 referencing, and then I send them to you and you will 14 have the definitive answer on the citizenship and who 15 they were. 16 REP. SWALWELL: 17 tell you what the citizenship was? 18 19 MR. WYLIE: I’m not 100 percent sure. REP. SWALWELL: 21 MR. WYLIE: European. But did anyone ever I believe I know, but I just -- 20 22 Okay. So I don’t want to misspeak. Okay. But I believe they were Eastern Page 116 1 REP. SWALWELL: 2 MR. WYLIE: Have you ever been to Russia? Have I -- no, I have never been to 3 Russia. 4 but I have never been to Russia proper. I’ve been to the Baltics and Eastern Europe, 5 REP. SWALWELL: Did you ever go to the Baltics 6 or Eastern Europe while you were working for Cambridge 7 Analytica? 8 MR. WYLIE: 9 REP. SWALWELL: Yes. Did you ever receive any -- or 10 were you ever subject to any unusual approaches or 11 anything that kind of just seemed unusual? 12 contact you that struck up conversation and just seemed 13 out of place? 14 MR. WYLIE: Did someone I could -- if I answered yes, if I 15 can just explain, that, yes, there were lots of various 16 approaches from people, you know, for example in 17 Lithuania about, you know, wanting to do business or 18 wanting to explore what it was that CA was doing or 19 what I was doing. 20 clear, I have never actually engaged in business there 21 personally or on behalf of Cambridge Analytica. 22 So, first of all, just to be super REP. SWALWELL: You’re saying this was by a Page 117 1 solicited -- 2 MR. WYLIE: But part of -- part of it might 3 just simply be a cultural difference in how people 4 approach things. 5 Region, and I have been approached by people who are 6 aware of the work that we were doing and were 7 interested in it. 8 9 But, yes, I have been to the Baltic REP. SWALWELL: And when these individuals made these approaches, did you have any reason to 10 believe that -- or had you made them aware of the work 11 you were doing, or were they just completely random 12 approaches and they told -- 13 MR. WYLIE: Well, I was -- 14 REP. SWALWELL: 15 MR. WYLIE: -- you that there -- -- I was visiting -- because 16 several of the other staff members at CA were from 17 Baltic Region or Eastern Europe, my understanding was 18 that they had spoken to people back in their respective 19 home countries about some of the work that we were 20 doing. 21 22 So whether people know it through just, you know, completely innocent conversation or know it -- in Page 118 1 other words, I don’t -- I don’t know. 2 REP. SWALWELL: Going back to Mr. Bannon and 3 the -- as you described it, discouragement of voters or 4 voting, you said you heard it on conference calls, what 5 exactly do you remember him saying as it related to 6 discouraging voter participation? 7 MR. WYLIE: Finding -- finding ways -- so my 8 recollection of it was that Steve Bannon was 9 particularly focused more -- just generally on African 10 American in terms of voter disengagement. 11 So my recollection of those calls was that he 12 -- he wanted to understand not just how to motivate his 13 base or his supporters, but how to demotivate his 14 opponents and discourage them from participating in an 15 election, because, in his mind, that’s the most -- like 16 that’s a key piece of winning. 17 up. 18 19 You push them down. REP. SWALWELL: You push your people It makes it easier to win. Was Mr. Bannon ever involved in developing tests or models for this type of -- 20 MR. WYLIE: He’s not a technical -- 21 REP. SWALWELL: 22 MR. WYLIE: -- engagement? He’s not a technical person. He Page 119 1 has a good general understanding of technology, but he 2 is not a developer. 3 handle data, but he was aware of what we were doing for 4 sure. 5 who authorized spending. 6 start a new project or a new research stream, they had 7 to authorize it so that there was appropriate funds to 8 spend on it. 9 So he didn’t directly model or And because he and Rebecca Mercer were the ones REP. SWALWELL: 10 MR. WYLIE: Anytime the company wanted to Who is Robert Murtfield? Robert Murtfield was an associate 11 of a barrister, a lawyer in the UK named John Jones. 12 John Jones was Julian Assange’s lawyer. 13 After -- John Jones had some kind of 14 psychological breakdown, after the Hillary Clinton 15 emails were released, he then walked in front of a 16 train and killed himself. 17 his assistant was Robert Murtfield, who then -- or his 18 clerk. 19 is technically different from his assistant, then was 20 hired by Cambridge Analytica and now -- and still works 21 for Cambridge Analytica. 22 REP. SWALWELL: Sorry. And his assistant, I -- yes, I’ve been corrected, his clerk, which Did you know -- do you know Page 120 1 Mr. Murtfield? 2 3 MR. WYLIE: personally, no. 4 5 I don’t -- I don’t know him REP. SWALWELL: Did you ever communicate with him? 6 MR. WYLIE: I believe I’ve met him, but I -- I 7 didn’t have any -- I knew who he was and -- but I 8 didn’t have any like direct working relationship with 9 him. 10 REP. SWALWELL: To your knowledge, did Julian 11 Assange have any relationship with Cambridge Analytica, 12 or do you suspect, based on some of the circumstantial 13 evidence you’ve pieced together, that Mr. Assange had a 14 relationship with Cambridge Analytica? 15 MR. WYLIE: So this is where working with the 16 British inquiry would be very helpful. 17 from various Cambridge Analytica employees talking 18 about the relationship with Julian Assange, that that 19 relationship happened -- was -- that that relationship 20 happened over a year before the American election 21 started, well over a year before the American election. 22 There are tapes So it wasn’t just the email in June, or Page 121 1 whatever the official -- and I’ve also seen emails that 2 John -- before he died, that John Jones was -- because 3 John Jones also, in addition to being Julian Assange’s 4 lawyer, also did work for Cambridge Analytica in 5 various ways. 6 He wrote -- I think he even went and lobbied 7 in the United States for some of their clients, I 8 believe, or wrote articles and things like that. 9 so I have seen, although I don’t possess, but I know 10 And where they are, emails that allude to WikiLeaks. 11 REP. SWALWELL: And so just based on what 12 you’ve laid out for us, you know of Russians working 13 for Cambridge Analytica who traveled to Russia. 14 know of individuals working for Cambridge Analytica who 15 traveled to Russia. 16 doing work on how Russia is viewed in the United 17 States. 18 Analytica, a year before the hacking of Democratic 19 emails, were in contact with Julian Assange. 20 all accurate? 21 22 You You know of Cambridge Analytica And you know that individuals at Cambridge MR. WYLIE: Yes. Is that And in addition to that, I also know that Cambridge Analytica was advertising the Page 122 1 projects that they were working on in the United States 2 to LUKOIL and the senior executives at LUKOIL. 3 REP. SWALWELL: Were you aware of, once you 4 left Cambridge Analytica, but in your continuous 5 contacts with friends there, of any effort to obtain 6 Hillary Clinton’s quote/unquote “deleted emails”? 7 MR. WYLIE: Yes. So some of the tape 8 recordings reference that. Something else to just 9 point out is that the firm SCL has a history of making 10 use of hacked material. 11 the company utilized medical records and private emails 12 of the current president of Nigeria in trying to create 13 Kompromat. 14 So, for example, in Nigeria, The person who organized that project in 15 Nigeria, Brittney Kaiser, also had a very significant 16 relationship with John Jones before he did. 17 believe he was supervising her -- her master’s degree, 18 or something like that, and then she then went on and 19 worked at his firm before she went to Cambridge 20 Analytica. 21 22 REP. SWALWELL: As I So what were you hearing from colleagues of yours at Cambridge Analytica? What Page 123 1 personal knowledge do you have of discussions about 2 obtaining Hillary Clinton’s deleted emails? 3 MR. WYLIE: I don’t have personal knowledge 4 because I never discussed access using hacked material 5 in the United States, or anywhere, actually, for that 6 matter. 7 illegal, that’s a crime. 8 9 That’s a line that I wouldn’t -- that’s But I have seen email documentation, tape recordings, and also have spoken to various employees 10 who also, you know, reference relationships with 11 WikiLeaks and, in particular, acquiring Kompromat in 12 the same way that they would attempt to obtain 13 Kompromat in an African country. 14 15 16 17 18 19 20 21 22 REP. SWALWELL: Who were they trying to obtain Kompromat on? MR. WYLIE: The Democrats. So Hillary Clinton, in particular. REP. SWALWELL: And is this publicly available information, or is this something that you are -MR. WYLIE: Some is and some isn’t. in the possession of the inquiry in Britain. REP. SWALWELL: Some is It’s -- And who are the individuals Page 124 1 seeking the deleted emails? 2 Analytica employees? 3 MR. WYLIE: Yeah. 4 then I believe Nigel Oakes. 5 is on tape talking about it. 6 REP. SWALWELL: Are these Cambridge Alexander Nix and I -- and Nigel Oakes, for example, Now, have you seen the 7 publicly released email where I believe Mr. Nix is 8 talking with an individual about a conversation or an 9 approach he had made to Julian Assange? 10 MR. WYLIE: Yeah. 11 REP. SCHIFF: 12 REP. SWALWELL: Rebekah Mercer. Yeah. Rebekah Mercer. Are 13 you saying that, aside from this, there were other 14 efforts that you’re aware of to obtain -- 15 MR. WYLIE: That’s my understanding. So I 16 think if you -- I can -- if you don’t already have a 17 relationship with Damian Collins, who is the chair of 18 the DCMS Committee at the UK Parliament, I am happy to 19 -- I know him very well, so I can introduce you, 20 because this is something that they are looking at. 21 And then I can also speak to others who have 22 documentation around this area. Page 125 1 2 REP. SWALWELL: Chairman. 3 4 I’ll yield back to you, Mr. REP. SCHIFF: Thank you. Thank you, Mr. Swalwell. 5 Just to follow up on that, in the public email 6 or public reported email, Rebecca Mercer suggests that 7 Cambridge Analytica reach out to Julian Assange and try 8 to basically house the -- and archive the emails. 9 Nix, I think, reportedly responds that he tried that. 10 And So you’re aware of emails and tape recordings 11 that indicate that it was more than Nix simply asking 12 his assistant to reach out and being rebuffed? 13 MR. WYLIE: 14 REP. SCHIFF: Yes. Do you know whether Cambridge 15 Analytica had any success in getting information for 16 WikiLeaks? 17 MR. WYLIE: That I don’t -- I’m not definitive 18 on. 19 pieces of evidence that -- that show a wider 20 relationship with -- or association with Julian Assange 21 and WikiLeaks that go back much further than just that 22 brief moment that -- or at least the narrative that it What I can help you get access to is various Page 126 1 was only just an email or just a random cold call to 2 WikiLeaks. 3 REP. SCHIFF: So when Mr. Nix purportedly 4 asked his assistant to reach out to Julian Assange, it 5 wouldn’t have been a cold call. 6 relationships there. 7 MR. WYLIE: 8 9 10 There were Can I tell you why that doesn’t make sense? REP. SCHIFF: MR. WYLIE: Yeah. So Brittney Kaiser, who after the 11 stories came out she left the company but was there 12 until a couple of weeks ago -- she was the business 13 development director -- she had worked for John Jones, 14 so Julian Assange’s lawyer. 15 Robert Murtfield coming on, she also came on. 16 And so in addition to And the reason why I think that it doesn’t 17 make sense to just cold call is that she would have had 18 very easy access to WikiLeaks and to Julian Assange 19 because she worked for his now-deceased lawyer, who had 20 a relatively close relationship with Julian Assange. 21 And so for me to go -- and it doesn’t make 22 sense that he would just cold call when he had a very - Page 127 1 - when he would have had a much warmer reception using 2 a network. 3 4 REP. SCHIFF: left WikiLeaks and came to work at Cambridge Analytica. 5 MR. WYLIE: 6 WikiLeaks. 7 and -REP. SCHIFF: 9 MR. WYLIE: 11 He was working -- he didn’t leave He was also working for Cambridge Analytica 8 10 So the Julian Assange lawyer The same --- and then he committed suicide. REP. SCHIFF: So the same lawyer worked for both Julian Assange and Cambridge Analytica. 12 MR. WYLIE: Yeah. 13 MS. ALLEN: Is it helpful to explain the split 14 profession, the barrister -- the role of a barrister as 15 opposed to what he was doing -- 16 MR. WYLIE: If you think it’s relevant. 17 MS. ALLEN: Well, I’m not sure that it is, but 18 19 20 the -- as a barrister -REP. SCHIFF: Well, they were both clients of the lawyer. 21 MR. WYLIE: 22 REP. SCHIFF: Yeah. So the lawyer wasn’t in-house Page 128 1 with WikiLeaks or in-house with Cambridge Analytica. 2 MR. WYLIE: 3 REP. SCHIFF: 4 MR. WYLIE: 5 barristers are never employed. 6 independent. REP. SCHIFF: 8 MR. WYLIE: 10 He had both of these clients. Yes. Barristers are always -They are always It’s just -- 7 9 Yeah. Oh, I see. -- a function -- it’s just how the -- there’s two types of lawyers in the UK. There’s solicitors -- 11 REP. SCHIFF: 12 MR. WYLIE: And the barristers --- who can be employed, and 13 barristers are -- do you want -- you’re the lawyer. 14 You can explain it. 15 REP. SCHIFF: 16 MS. ALLEN: No, I think I understand. Yes. So this is how the client 17 relationship -- the main -- the primary client 18 relationship, and they then instruct the barrister. 19 John Jones was a barrister. 20 lawyer who had instructed John Jones, unless he was 21 working outside his normal chambers to give informal 22 advice. So there would have been a Page 129 1 2 REP. SCHIFF: clients of the same barrister as -- 3 MR. WYLIE: 4 REP. SCHIFF: 5 So they were two different Yes. -- who ended up taking his life. And that barrister’s clerk also -- 6 MR. WYLIE: Robert Murtfield, who -- 7 MS. ALLEN: And as a clerk, the clerk is 8 attached to the chambers, which is the place which 9 houses a group of barristers. It’s not a -- it’s not a 10 law firm as such. 11 people, and they have a -- gather together onto the 12 banner of a chambers, and the chambers employs the 13 clerk. 14 15 It’s a group of self-employed So the clerk would have -MR. WYLIE: But he was the clerk that worked MS. ALLEN: Yeah. for -- 16 And they’re assigned some 17 barristers, but he would have worked for others as 18 well. 19 MR. WYLIE: Yeah. And so he -- so to put it 20 in just simple terms, we have a lawyer who was working 21 for WikiLeaks and working for Cambridge Analytica. 22 When he died, his clerk, assistant, person, then went Page 130 1 over to Cambridge Analytica and -- and then also 2 Brittney Kaiser, who wasn’t a direct employee of his, 3 but she was supervised by him for her academic research 4 and then she went on and worked at the same chambers, 5 which is kind of like a law firm. 6 She then also went to Cambridge Analytica, and 7 she was the one who, for example, introduced the 8 Cambridge Analytica to a firm in Nigeria or a foreign 9 firm that accessed somehow the private medical records 10 of the now-president of Nigeria. 11 She was the one who organized that 12 relationship and that project, which likely involved 13 hacking. 14 interested in speaking to. So she is probably somebody that you might be 15 REP. SCHIFF: 16 MR. WYLIE: 17 REP. SCHIFF: 18 MR. WYLIE: 19 REP. SCHIFF: And tell me her name again. Brittney Kaiser. So Brittany Kaiser -She is an American citizen. -- worked at one project for 20 Cambridge Analytica in which they had records for use 21 in a political campaign in Nigeria. 22 MR. WYLIE: Is that -- So projects that involved -- I’m Page 131 1 just conscious of that this might be published on 2 record and it doesn’t have privilege. 3 misappropriated private or personal information of 4 opposition politicians in Nigeria and several Caribbean 5 countries where she was involved in managing the 6 relationships of the contractors who, through various 7 means, were able to acquire that information. 8 9 REP. SCHIFF: So So that was part of her responsibility at Cambridge Analytica, or was it -- 10 MR. WYLIE: I don’t know if it was a formal 11 job -- I don’t think she would have put it on her 12 LinkedIn, that that’s something that was her formal job 13 role. 14 and there are several instances of her involving in 15 that kind of work. But it is something that she was involved in, 16 17 REP. SCHIFF: And she was working for the clerk for the barrister who had -- 18 MR. WYLIE: She was working for the chambers 19 of the barrister. She was working with the barrister 20 who was -- who represented Julian Assange at the 21 chambers, and then also he supervised her master’s 22 degree. Page 132 1 2 REP. SCHIFF: Okay. And then she leaves that chambers to go to work at -- 3 MR. WYLIE: For Cambridge Analytica. 4 REP. SCHIFF: Okay. Okay. So you have 5 someone with experience in using misappropriated 6 materials through Cambridge Analytica to influence an 7 election, who has a relationship with people who have a 8 relationship with Julian Assange. 9 MR. WYLIE: 10 Yes. REP. SCHIFF: And there are documents, 11 recordings about an approach to try to obtain those 12 materials from Julian Assange, but you don’t know 13 whether they were actually acquired or used by 14 Cambridge Analytica. 15 MR. WYLIE: 16 That’s the one thing that I can’t confirm to you. 17 REP. SCHIFF: Okay. Do you know whether -- 18 and we would love to speak to our counterpart in 19 Britain. 20 that? 21 22 Do you know whether they have evidence of MR. WYLIE: So they have a lot of the recordings that I’m referring to. I actually helped Page 133 1 them get the recordings, so I wasn’t in possession, but 2 I facilitated them acquiring them. 3 know the individuals who have some of the emails and 4 documents in relation to that. 5 And then I also I also know that there are several journalists 6 who have some material who may, in some way, be willing 7 to show them. 8 9 10 REP. SCHIFF: And you mentioned that Nigel Oakes was involved in this effort. at SCL? What was his role I know he was one of the original -- 11 MR. WYLIE: He is -- yeah, he goes way back in 12 the firm. 13 He is -- there are recordings of him talking about 14 WikiLeaks. 15 16 17 So he is one of the shareholders, directors. So he knows about the situation. REP. SCHIFF: And he is still an investor in 2016? MR. WYLIE: He is, I believe, currently still 18 a major shareholder in SCL, not Cambridge Analytica 19 proper, but in SCL Group. 20 REP. SCHIFF: Did you ever come across Roger 21 Stone while you were at Cambridge Analytica or learn of 22 any interaction that he had with any of the principals Page 134 1 there? 2 MR. WYLIE: 3 REP. SCHIFF: 4 Okay. I have some other questions, but I’m going to go to Mr. Castro. 5 6 No, not while -- no. REP. CASTRO: testimony today. Sure. Thank you for your You’re American? 7 MR. WYLIE: I’m Canadian, actually. 8 REP. CASTRO: 9 MR. WYLIE: Oh, Canadian. But I have residency in the UK. 10 I’ve lived in the UK since 2010. 11 REP. CASTRO: Okay. Do you know whether any 12 other Americans were involved with Cambridge Analytica 13 in the activities that you’ve been describing here in 14 this interview? 15 16 MR. WYLIE: citizen. So Brittney Kaiser is an American She lives in New York. 17 REP. CASTRO: 18 MR. WYLIE: 19 20 21 22 And what was her role again? She was director of business development, I believe was her title. REP. CASTRO: Is she the one that had a role in Nigeria and the Caribbean nation? MR. WYLIE: And Caribbean, yeah. And also, Page 135 1 she is the one that worked with John Jones, the Assange 2 lawyer. 3 4 Other Americans who were involved -- Steve Bannon. 5 REP. SCHIFF: 6 MR. WYLIE: 7 REP. SCHIFF: 8 Mike Flynn was as well. Yes. Can you tell us about his involvement? 9 MR. WYLIE: I don’t know that much about what 10 exactly he was doing, but I know that he was an 11 advisor, and one of the roles that he had was to go and 12 solicit clients. 13 the firm made a very concerted effort to make it known 14 that they had these connections with the Trump 15 administration, that they got Trump elected, and they 16 were looking to, you know, solicit clients through -- 17 because of that. 18 I know that after Trump got elected, I know that Alexander Nix, after Trump got 19 elected, made a concerted effort to try to meet as many 20 people as he could. 21 the foreign commonwealth office, for example, in the 22 UK. I know that he had a meeting with Page 136 1 I believe he met with Boris Johnson, so he -- 2 so he and the firm went around essentially advertising 3 this relationship, and I think with the view of trying 4 to profit off of it. My understanding of what Michael 5 Flynn’s role was was: a) as an advisor to the firm, 6 but b) to use his network of contacts and then also 7 this sort of Trump brand to solicit clients for the 8 company. 9 REP. CASTRO: And do you know, who did the 10 company share data with? 11 groups that you all were swapping data with at all or 12 coordinating with? 13 MR. WYLIE: Were there data partners So in terms of acquiring data, 14 there were several partner companies. 15 them were just very ordinary, so not that interesting, 16 but I know that they used Experian data, another 17 company called Info Group. 18 standard. 19 Group data. 20 Just -- some of All of this is sort of very Some of your own companies might use Info I know that after Brittney Kaiser came onto 21 the company, they started exploring relationships with 22 the Ergen family, so Chase Ergen, who my understanding Page 137 1 is that they -- that this family is, if not 2 shareholders, exert some sort of significant influence 3 over Dish Network in the United States. 4 5 6 REP. CASTRO: Do you know what kind of information or data was being shared with them or -MR. WYLIE: I don’t know in particular. I 7 know that this was something that they were quite 8 interested in. 9 with a lot of the various companies in, for example, 10 I know that the company’s relationship Mexico, like Pig.gi, which is a data harvesting app -- 11 REP. CASTRO: 12 MR. WYLIE: What is it called? Pig.gi, P-i-g dot g-i. 13 Mowasat, which is also in Mexico. 14 REP. CASTRO: 15 MR. WYLIE: Or How do you spell that one? M-o-w-a-s-a-t. Both of those 16 companies, the Ergens are involved in. The Ergens were 17 also involved in, I believe, Nigeria and several other 18 countries. 19 understanding is that the projects that they were doing 20 -- that they are doing in Mexico are sort of research 21 and development projects with different means of data 22 harvesting, where oversight is not necessarily robust, The various sort of projects -- my Page 138 1 to see what they can do and what they can get away 2 with, to then export that to other parts of the world 3 that they -- that they operate in. 4 Perhaps what I can do is I can give you -- I 5 can come back to you guys and give you just a complete 6 list of every company or organization or individual 7 that I’m aware of where they had some sort of data 8 transaction. 9 REP. CASTRO: 10 MR. WYLIE: 11 REP. CASTRO: 12 MR. WYLIE: 13 REP. CASTRO: 14 That would be helpful. If that -- if that’s helpful. That would be great. Yeah. Yeah. Thank you. with any -- any Russians or anyone in Russia? 15 MR. WYLIE: So I’m not sure if you were here 16 when I was speaking about Russia before. 17 maybe -- 18 19 20 What about REP. CASTRO: So I’ll just If you’ve covered it, then that’s fine. MR. WYLIE: Okay. Just very quickly, Dr. 21 Kogan was working in Russia at the time. His research 22 at St. Petersburg University was on online trolling and Page 139 1 something called Dark Triad Traits, which is 2 Narcissism, Machiavellianism, and Psychopathy. 3 He was doing that at the same time that he was 4 managing the Facebook harvesting project, and in 5 addition to that happening, which, by the way, SCL was 6 advertising to other clients that it had the 7 applications of this Russian work. 8 9 It was also interfacing with LUKOIL, so advertising. I can perhaps, if it’s -- 10 REP. SCHIFF: 11 MR. WYLIE: 12 REP. CASTRO: We covered Brad Parscale also? 13 REP. SCHIFF: We didn’t really get into Brad REP. CASTRO: What was the relationship, if 14 Yeah. Parscale. 15 16 We’ll go back to -- any, or your interfacing with Brad Parscale? 17 MR. WYLIE: 18 left, so I didn’t -- 19 REP. CASTRO: I didn’t -- that was after I had So did you hear anything about 20 it, or it was coming, it was on the horizon, or 21 anything? 22 MR. WYLIE: No, not with respect to him. No. Page 140 1 REP. CASTRO: Okay. 2 REP. SCHIFF: So I have kind of a lightning 3 round, a number of questions. 4 MR. WYLIE: 5 REP. SCHIFF: 6 Sure. order. 7 MR. WYLIE: 8 REP. SCHIFF: 9 So there will be no particular Nix. Okay. We had an interview with Mr. If he testified that Cambridge Analytica received 10 no data through Facebook, would that be an inaccurate 11 or misleading statement? 12 MR. WYLIE: So he also said that to the 13 British Parliament. 14 Parliament, I will say the exact same thing to you. 15 Facebook data -- the company’s foundational models were 16 based on Facebook data. 17 And as I said at the British I think it’s incredibly misleading to say that 18 no Facebook data was used to develop the algorithms or 19 the targeting or the -- or to conduct the message 20 testing of the company. 21 22 The firm -- and I can provide you with these documents. It has Alexander’s signature on it. The Page 141 1 firm engaged Dr. Kogan specifically because his 2 application harvested Facebook data, and specifically 3 because it harvested not only the Facebook data of the 4 user but also their friend network. 5 contracts. 6 contracts have Alexander’s signature on it. 7 8 That is in project documentation. The So I don’t understand why he would tell you that they didn’t use Facebook data. 9 10 That is in REP. SCHIFF: Is that what they teach you at Eton? 11 MR. WYLIE: 12 REP. SCHIFF: 13 MR. WYLIE: 14 REP. SCHIFF: I -That’s a rhetorical question. Yeah. You mentioned that one of the 15 things that we should be considered about is potential 16 violations of the Foreign Corrupt Practices Act. 17 were you referring to there? 18 19 MR. WYLIE: so can I -- sorry. Yes. REP. SCHIFF: 21 MR. WYLIE: 22 So one -- so I am -- I am -- Can I ask a question? 20 What So -- Yes. -- if I speak about certain things, I -- so I understand that this will be made Page 142 1 public at some point. 2 REP. SCHIFF: Yes. And I want to -- I was 3 going to wait until the end of the interview to ask 4 you. 5 yesterday released a summary of key parts of your 6 testimony. The Judiciary Committee that you interviewed with 7 MR. WYLIE: 8 REP. SCHIFF: 9 10 Yeah. Including some excerpts. don’t know if they discussed that with you in advance or not. 11 MR. WYLIE: I don’t believe they -- they 12 didn’t discuss the excerpts. 13 going to publish. 14 record. I knew that they were I treated it as entirely on the 15 REP. SCHIFF: 16 MR. WYLIE: 17 I Okay. So -- So we didn’t go nearly as in-depth on certain topics. 18 REP. SCHIFF: What I thought we would do when 19 we finish is normally our interviews are in closed 20 session. 21 MR. WYLIE: 22 REP. SCHIFF: Right. Subject to our releasing the Page 143 1 transcript at a point when we decide to release it. 2 MR. WYLIE: 3 REP. SCHIFF: Okay. Now, as the majority is not 4 participating, we are not operating under the normal 5 rubric. 6 MR. WYLIE: 7 REP. SCHIFF: 8 So we have more flexibility than we would otherwise. 9 MR. WYLIE: 10 11 Right. REP. SCHIFF: Right. I don’t want to release something that is going to create an issue for you. 12 MR. WYLIE: 13 REP. SCHIFF: Sure. Sure. Without there being a good for 14 it, without giving you a heads-up, and anything along 15 those lines. 16 There were a couple of things that I wanted to 17 ask you about at the end whether you had any 18 reservation with our discussing publicly generally what 19 you had said without releasing the transcript. 20 I think -- 21 MR. WYLIE: 22 REP. SCHIFF: I mean, Okay. -- I think, for example, it will Page 144 1 be of great interest to people that there were more 2 connections to WikiLeaks than is publicly known. 3 MR. WYLIE: Right. 4 REP. SCHIFF: But I would want to ask you 5 about a couple of different areas and whether you are 6 comfortable with our sharing at least a 7 characterization, if not exact comments that you made. 8 MR. WYLIE: Sure. 9 REP. SCHIFF: But on this particular topic 10 that I asked about FCPA, if there is something that 11 concerns you, then, obviously, that will be something 12 we will discuss after the interview. 13 And down the road, if we’re at a point of 14 releasing the transcripts of interviews, we would want 15 to work with you and try to address any concerns you 16 have of timing, over any redactions that we need to 17 make. 18 MR. WYLIE: 19 REP. SCHIFF: 20 MR. WYLIE: 21 22 Sure. And that type of thing. Sure. So if we could have that conversation perhaps off record for -REP. SCHIFF: Yeah. Page 145 1 MR. WYLIE: 2 REP. SCHIFF: 3 MR. WYLIE: 4 -- for now, and then -Sure. -- simply because -- for two reasons. 5 REP. SCHIFF: 6 (Recess taken from 6:04 p.m. to 6:39 p.m.) 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Can we go off the record? Page 146 1 E V E N I N G 2 (6:39 p.m.) 3 REP. SCHIFF: 4 very short to these. 5 MR. WYLIE: 6 REP. SCHIFF: 7 If you could make your answers I will try to be succinct. If there are areas where I need to know more, I will then ask follow-up. 8 MR. WYLIE: 9 REP. SCHIFF: 10 S E S S I O N Okay. Tell me, Michael Kosinski, he worked with Kogan or -- 11 MR. WYLIE: He worked with Kogan initially. 12 He helped prepare some of the planning and he did some 13 of the preparatory work for the data harvesting 14 project. 15 He did not continue on that particular data 16 harvesting project -- I'm referencing the Facebook 17 project -- because he demanded half a million dollars 18 and 50 percent royalties, whereas Kogan said that he 19 wanted to do it for this academic institution and, 20 therefore, would do it at cost. 21 22 REP. SCHIFF: Did Kosinski have any relationship with any Russian institutions? Page 147 1 MR. WYLIE: Yeah, and also to be clear, he was 2 the deputy. 3 Director at the time of the Cambridge Psychometric 4 Center. 5 the equivalent type of harvesting. 6 identical capacity that Kogan had. 7 Kogan -- sorry -- Kosinski was the Deputy So he also had his own applications that did So he had the He had already amassed a very large data set, 8 mostly on Americans. 9 academic use. That originally was harvested for One of the things that I can provide you 10 is a screen shot of Kosinski saying that -- it's him 11 boarding a plane, and he says -- he makes reference to 12 the fact he's going to Russia, meeting with senior 13 politicians, including the Prime Minister of Russia, 14 where he was going to present some of the work that he 15 was doing, which was identical to what Kogan was doing. 16 17 18 19 20 Because originally, they were working together, and -REP. SCHIFF: And do you know whether he had formed any relationship with the Russians? MR. WYLIE: I don't know in terms of any 21 employment or any sort of that kind of relationship. 22 know that he did presentations in Russia to politicians I Page 148 1 about this work. 2 3 REP. SCHIFF: Do you know whether he shared his data? 4 MR. WYLIE: I don't know if he -- I don't know 5 if he shared data, but I do know that he did 6 presentations on that data where that data may have 7 been used. 8 9 10 REP. SCHIFF: And how voluminous was the data compared to what Kogan got? MR. WYLIE: My understanding was it was in the 11 millions, was my understanding, but the point that I'd 12 make is he had the same capacity to harvest data. 13 I'm not sure if he went and harvested more data or if - 14 - but the data set that he had, he had already built 15 algorithms, providing algorithms with, and had already 16 validated those algorithms. 17 research that was published presented that. 18 Indeed, some of his Some of it may be of interest to you. 19 that research was funded by DARPA. 20 can't answer that question, you know, directly. 21 22 REP. SCHIFF: So Some of So I'm not sure. I Well, in those papers, he would not have presented the algorithms. Page 149 1 2 MR. WYLIE: REP. SCHIFF: I see. Anything else about Kosinski we should know? 5 6 He just would have presented results. 3 4 No, no, no. MR. WYLIE: Not -- not -- no, I don't think there's anything in particular. 7 REP. SCHIFF: There's a New York Times article 8 that said two former company insiders, I assume 9 referring to Cambridge Analytica, said LUKOIL was 10 interested in data targeting American voters. 11 know who the other of these two insiders refers to? 12 MR. WYLIE: Do you I have an idea, but I can't say 13 definitively one way or the other, but there were 14 several. 15 aware of the LUKOIL project. There were -- I wasn't the only one who was 16 REP. SCHIFF: 17 MR. WYLIE: 18 21 22 Although I was the one that had the actual documentation on it. 19 20 Okay. REP. SCHIFF: What role, if any, did Dr. Kogan in LUKOIL? MR. WYLIE: I don't think that Kogan actually met with LUKOIL, but I do know that Alexander made it Page 150 1 known to LUKOIL -- 2 REP. SCHIFF: 3 MR. WYLIE: That Kogan was --- that Kogan was one of -- that 4 we had a Russian in Russia who was working on it, and 5 you're a Russian company, and this is something that 6 you would like. 7 And so I don't know if Kogan actually met with 8 LUKOIL, but I do know that Alexander Nix mentions and 9 referred to. 10 11 REP. SCHIFF: What do you know about John Bolton's PAC? 12 MR. WYLIE: Yeah, John Bolton's PAC was a 13 client. 14 Cambridge Analytica actually. 15 that they funded made use of models that were built 16 with the Facebook data. 17 They were one of the first clients of They used -- the project My understanding of the project was that John 18 Bolton wanted to explore ways of, for lack of a better 19 term, increasing sort of militaristic views in 20 Americans; that they had their sort of hypothesis or 21 theory, was that Americans are becoming too weak on a 22 sort of military intervention, and that they wanted to Page 151 1 explore ways of making people more supportive of robust 2 military intervention around the world. 3 REP. SCHIFF: 4 Cambridge Analytica? 5 MR. WYLIE: And did the Bolton PAC retain Were they a paying customer? Yeah, yeah, and I can actually 6 provide you with -- I believe I can provide you with 7 the contracting and the project specification. 8 9 10 REP. SCHIFF: And there were specific races that Bolton's PAC targeted that -MR. WYLIE: I believe there was a combination 11 of states and districts of interest, but also more 12 generally research on this sort of strength, this sort 13 of militarism. 14 Americans more militaristic? 15 16 17 How can we just generally make REP. SCHIFF: Okay. What was Palantir's relationship to Cambridge Analytica? MR. WYLIE: So one of the -- so before I 18 joined the company, and by "the company," I mean SCL 19 Group, Sophie Schmidt works for SCL Group. 20 Schmidt is the daughter of Eric Schmidt, the Chairman 21 of either Google or Alphabet. 22 Sophie She encouraged Alexander Nix to get in touch Page 152 1 with Palantir. 2 first introductions to Palantir. 3 I believe that she facilitated the When I joined, one of the first emails that I 4 got was asking about my views on Palantir and whether 5 we should engage. 6 Alexander Nix, he wanted to know what we should -- that 7 he had met people at Palantir, and he wanted to pursue 8 a relationship with them. 9 One of the first meetings I had with I told him, I said, "Well, that might make 10 sense for some of the, you know, military or counter- 11 extremism projects that, you know, that the company 12 works on, but they didn't" -- I literally, and I think 13 I even put this in an email -- that they didn't make 14 sense for a campaign, any political projects. 15 Nonetheless he pursued that relationship. 16 had several meetings at Palantir. 17 meetings at Palantir in their European office. 18 that's based in London, in Soho. 19 He I got invited to So So after Cambridge Analytica got set up, so in 20 the spring of 2014, Palantir staff started showing up 21 in the office. 22 very regularly. One staff member in particular came in I was told that they were helping Page 153 1 advise on some difficult -- I don't recall the 2 specifics of it, but some kind of difficult problem 3 that the data science team was experiencing with 4 algorithms or some kind of complex algorithmic 5 structure that they were looking at, and that also the 6 staff members were just generally interested, in 7 particular, as to what kinds of projects were we 8 working on and to see what could be done with the data 9 that we were acquiring, including the Facebook data. 10 So I have emails from some of these staff 11 members talking about Facebook data acquisition, ideas 12 for modeling it, ideas for acquiring it. 13 One of the things that I found slightly 14 disconcerting about that, which is why it sticks in my 15 head, was that we were all instructed that when we were 16 dealing with anybody from Palantir on the staff list, 17 they had pseudonyms. 18 there was at least one time where I saw a staff member 19 being paid in cash. 20 was. 21 22 So they had fake names, and that I'm not sure exactly what that It was slightly unusual as to why they kept coming in. Palantir says that they were doing it in a Page 154 1 personal capacity. 2 that is unusual about that is that it was during 3 business hours, and it was regular. 4 somebody coming in for an hour in the evening or on the 5 weekend just to give some casual advice. 6 were regular intervals of meetings and actual work that 7 was being done. 8 9 The only thing that I would say So it wasn't like Like there But I don't know exactly what happened with that or what kind of data might have been taken, if 10 there was any data taken, but I know that there was one 11 Palantir staff member in particular who had access to 12 the Facebook data and was quite interested in it. 13 But I can't say exactly -- 14 REP. SCHIFF: So it wasn't clear why they were 15 there or what they were doing, and they had these 16 pseudonyms? 17 MR. WYLIE: Yes. So some of the emails I have 18 are from the personal accounts, personal email 19 accounts, of staff members, and then some of the emails 20 I have come from pseudonyms, but some of the emails 21 they signed off on their actual name and forgot that 22 they were using a pseudonym. Page 155 1 It might be that they were there in a personal 2 capacity, and that -- I don't know -- Palantir just has 3 very loose requirements as to coming into the office 4 and taking personal days regularly, but the thing that 5 I would say is that if you're doing something in a 6 personal capacity during the week and actually doing 7 some fairly complex work, it doesn't seem like a 8 personal project. 9 REP. SCHIFF: But any guess as to what the 10 deal was? 11 nationals working on a British campaign or any reason? 12 I mean, is there any rational expression for 13 Was this an effort to avoid foreign why all of the skullduggery? 14 MR. WYLIE: So I don't -- I shouldn't 15 speculate. 16 tell you sort of what I observed and what I know. I don't think I should speculate. 17 REP. SCHIFF: 18 MR. WYLIE: 19 REP. SCHIFF: So I'll Okay. Without speculating. What's your thought in terms of 20 how much Facebook knew about what was being done with 21 its data in real time? 22 MR. WYLIE: So Facebook has said that they Page 156 1 didn't know what was going on, but I remember at least 2 one instance where there was a delay in the next 3 tranche of data being acquired by Kogan and remember 4 talking to Kogan and just asking him what the delay was 5 about. 6 And he said something about he was having some 7 problem with the app, that Facebook either temporarily 8 shut it down or throttled the amounts of data he was 9 being to pull per minute from the app, but that he had 10 a conversation with somebody at Facebook. 11 was like an engineer at Facebook, and that then they 12 sort of turned the permissions back on or did something 13 that -- 14 REP. SCHIFF: I think it That would indicate though that 15 they were concerned about the volume of data being 16 acquired, but any indication that they knew that the 17 data was not being used for academic purposes, that, in 18 fact, it was being used by a commercial entity? 19 MR. WYLIE: That I don't know specifically, 20 but one of the other things that I should point out is 21 that Kogan's business partner, Joseph Chancellor, then 22 went and worked at Facebook. He still works at Page 157 1 Facebook. 2 And I believe they actually went and presented 3 the project to Facebook or learning from the project or 4 something. 5 to Facebook and actually presented something about the 6 projects. 7 I think there was something that they went This is something that I remember. I don't 8 know the exact details of it, but my recollection is 9 that they were relatively open or at least that 10 Facebook knew that there was this data that was being 11 acquired. 12 I don't know if Facebook knew specifically 13 whether it was commercialized or not. 14 that they weren't aware that it was commercialized; 15 that Facebook told me that they were told by Kogan that 16 it was for academic purposes. 17 REP. SCHIFF: They told me In Brittney Kaiser's testimony 18 before Parliament, she suggested that there were 19 additional Facebook-linked questionnaires and data sets 20 used by Cambridge Analytica. 21 MR. WYLIE: 22 REP. SCHIFF: Yes. And that the quantity of data Page 158 1 might be far greater. 2 MR. WYLIE: 3 REP. SCHIFF: Yeah. Is that a reference to 4 Kosinski's work and the data he was acquiring or what 5 do you think it refers to? 6 MR. WYLIE: No. So the purpose of Kogan was 7 -- so if you recall, after the money was deposited into 8 Cambridge Analytica, Steve Bannon then set this 9 deadline for September or around September to finish 10 data acquisition models, message testing, everything. 11 So that there was some time in between that and the 12 midterms that they could then play with it on some 13 campaigns. 14 So the original purpose of Kogan was sort of 15 this stopgap measure, which was we need to get lots of 16 data really quickly because Steve Bannon is demanding 17 it, and this scales really quickly. 18 But there was a decision made at the company 19 that essentially it's not smart business practice to -- 20 if data is the foundation of your company, it's not 21 good business practice to rely on an external vendor 22 solely to provide you that, and that internal capacity Page 159 1 should be built to create data acquisition in a sense. 2 So while Kogan was doing his harvesting 3 project, there were staff members at Cambridge 4 Analytica who were then exploring what kinds of apps or 5 browser extensions or various widgets or things to put 6 on people's computers that could replicate some of the 7 functionality that Kogan had, but it would be in-house 8 so they wouldn't have to rely on him. 9 there, that was a project. 10 11 So what Brittney Kaiser is referring to is the finished product of that exercise. 12 13 Even when I was REP. SCHIFF: And do we know how much data that was able to scrape? 14 MR. WYLIE: I don't know for certain, but if 15 they used applications that were developed outside of 16 Facebook, I mean, theoretically they could get tens of 17 millions of records that way. 18 19 REP. SCHIFF: Did you ever see any connection between Cambridge Analytica and Paul Manafort? 20 MR. WYLIE: Yes and no. So I know that Mark 21 Block -- so Mark Block -- are you familiar with Mark 22 Block? Page 160 1 Okay. So Mark Block was one of the people who 2 originally introduced Alexander Nix to Steve Bannon. 3 So he used to work for Americans for Prosperity. 4 was quite senior in that. 5 with the Mercers and Steve Bannon. 6 He Then he went to go and work He joined Cambridge Analytica after it got set 7 up, and one of his roles was to essentially just sort 8 of culturally validate this foreign firm with his 9 contacts in his particular wing of the Republican 10 Party. 11 So he didn't necessarily do anything in 12 particular aside from go and meet people and say, 13 "These guys are great and they're on our side, and you 14 know, I'm just like you. 15 They're just like us." But I know that he has or at least had a 16 relationship with Paul Manafort, and I believe that he 17 had meetings with Paul Manafort around the time that 18 Alexander Nix was also having meetings with Cory 19 Lewandowski. 20 But I don't believe that Mark Block still 21 works for Cambridge Analytica, but I know that he had 22 some relationship with Paul Manafort because I actually Page 161 1 knew who Paul Manafort was and who Roger Stone was 2 before they kind of became public figures after sort of 3 the Trump campaign, and it was because they -- I've 4 never met them, but their names would pop up on 5 occasion. 6 7 8 9 REP. SCHIFF: WikiLeaks or Julian Assange? MR. WYLIE: REP. SCHIFF: 11 MR. WYLIE: 12 REP. SCHIFF: 14 Not to my recollection. To me personally. 10 13 Did Steve Bannon ever talk about Yeah. Yeah. Did you ever hear of him discussing that with anyone else? MR. WYLIE: Not to my recollection, but he may 15 be referenced on some of these tapes that Damian 16 Collins has, but I don't -- I'm not sure on that. 17 REP. SCHIFF: I think you testified before the 18 Parliament about the ProtonMail accounts that Mr. Nix 19 utilized or others. 20 MR. WYLIE: 21 REP. SCHIFF: 22 Un-huh. Did you ever have any reason to believe that that was used to hide the various work Page 162 1 being done at Cambridge Analytica? 2 MR. WYLIE: So I'll just start by qualifying 3 my response, which is that there are many legitimate 4 reasons why you may use ProtonMail or self-destructing 5 emails. 6 country where you are being surveilled, which is a lot 7 of countries that SCL works in, or where evidence not 8 necessarily on your side, but on the client's side may 9 be discoverable and that would put them in jeopardy. 10 In particular, if you are operating in a That's not a reason why you would use it, but 11 there were instances where I think that Alexander may 12 have had relationships with people or conducted 13 business dealings or authorized something that could 14 have been unlawful and, therefore, it would have -- 15 that using technologies like that may have been useful. 16 17 REP. SCHIFF: Any specific occasions though where you know it was used for that purpose? 18 MR. WYLIE: 19 somebody of a crime. If I'm -- I'm on record accusing Is that -- 20 (Counsel conferred with the witness.) 21 MR. WYLIE: 22 Yeah, so if it's okay with you, I perhaps can provide you with some of the documentation Page 163 1 on some of the things that I have sent to the 2 authorities. 3 REP. SCHIFF: 4 MR. WYLIE: Yes. Because there are certain things 5 that may be potentially unlawful, but haven't been yet 6 -- haven't actually gone to court yet. 7 say that they are unlawful potentially is -- 8 9 10 REP. SCHIFF: then. So for me to Well, let me ask it this way Are you aware of use of Proton accounts to hide things that were unethical and potentially unlawful? 11 MR. WYLIE: I'm aware of people at Cambridge 12 Analytica or SCL Group using technologies like 13 ProtonMail or simply going and deleting things that may 14 have been potential evidence that would have been 15 relevant to an investigation. 16 17 REP. SCHIFF: Would you rather provide that to us in documentary form? 18 MR. WYLIE: 19 REP. SCHIFF: Yes. Okay. My staff may have a few 20 more questions for you, but is there anything I haven't 21 asked you about that you think that I should be aware 22 of? Page 164 1 MR. WYLIE: (Pause.) Sorry. I'm just -- I'm 2 just kind of going through our conversation just to 3 make sure. 4 REP. SCHIFF: 5 MR. WYLIE: Yes, sure. Not off the top of my head, but I 6 know five minutes after I leave there's going to be 7 something that I -- 8 9 REP. SCHIFF: We’ll follow up with you on the documents that you've offered to help with as well. 10 MR. WYLIE: 11 REP. SCHIFF: Yeah. So let me ask you this. I think 12 in terms of all of this is of great interest to us, and 13 very, very helpful. 14 Really grateful. I think what is probably of most immediate 15 interest, apart from the things that we won't talk 16 about, are the fact that there were a number of 17 connections between Cambridge Analytica and WikiLeaks. 18 MR. WYLIE: 19 REP. SCHIFF: 20 Un-huh. And the other is some of the detail on LUKOIL. 21 MR. WYLIE: 22 REP. SCHIFF: Okay. And the fact that LUKOIL was Page 165 1 showing an interest in these data sets involving U.S. 2 voters, and that LUKOIL operates sometimes as an 3 extension of Russian intelligence in places where the 4 Russian government can't operate, and the fact that 5 there were a number of people, including Dr. Kogan, who 6 were making frequent trips to Russia and were promoting 7 the work they were doing in terms of data analytics and 8 the type of modeling that Cambridge Analytica was using 9 in the United States. 10 11 I think those two areas would be of great interest. 12 MR. WYLIE: 13 REP. SCHIFF: 14 15 Okay. Do you have any reservations about my saying that we discussed this and -MR. WYLIE: No. I think one of the things 16 that I'd love to facilitate is connecting you with 17 Damian Collins and the investigation, the parliamentary 18 investigation that's happening because they have 19 certain documents or recordings or things that I think 20 might be interesting to you. 21 22 And also, given that a lot of the people involved in this are actually in the U.K., he might be Page 166 1 in a position if he is made aware or I know of 2 something, he might be in a position to compel some of 3 that evidence and then pass it to you. 4 5 MS. ALLEN: So -- He might be a source here potentially. 6 MR. WYLIE: Yes. I'm already doing it with 7 the Canadian Parliament because of Aggregate IQ, which 8 is the Canadian entity that was set up, which received 9 40 percent of all pro-Brexit spending. 10 But because they've been hiding from or not 11 necessarily hiding; because they've been not 12 forthcoming with information to the British 13 authorities, I connected the British committee and the 14 Canadian committee. 15 compelled the evidence that the British committee 16 wanted. The Canadian committee went and 17 REP. SCHIFF: 18 MR. WYLIE: 19 REP. SCHIFF: 20 21 22 Super. So potentially that could be -That would be very helpful. Thank you. And just, you know, so I'm clear, those two areas in terms of your testimony, you're okay with our Page 167 1 making public either in summary form or in transcript 2 form? 3 MR. WYLIE: If you're publishing transcripts, 4 would it be all right with you if we just have a chat 5 about what those transcripts look like? 6 REP. SCHIFF: 7 MR. WYLIE: 8 9 10 Certainly. Summary is totally fine, absolutely. REP. SCHIFF: MR. WYLIE: Yes. Just to make sure that everything 11 -- I want to make sure that things that are published 12 are absolutely solidly correct, and so if I realize 13 that I -- I don't think I have. 14 but I just want to make sure -- 15 REP. SCHIFF: 16 MR. WYLIE: 17 I haven't intended to, Sure. -- given the potential import. just want to make sure that everything is -- 18 REP. SCHIFF: 19 MR. WYLIE: 20 REP. SCHIFF: 21 MR. WYLIE: 22 REP. SCHIFF: We would be happy to. -- as accurate as possible. We would be happy to. Okay. Thank you very much. I Page 168 1 I'm going to leave you with my staff. 2 MR. WYLIE: 3 REP. SCHIFF: Sure. I think they only have a small 4 number of questions to go over. 5 end of the road. 6 MR. WYLIE: 7 REP. SCHIFF: So we're almost at the Sure. But thank you so much for 8 speaking out and agreeing to sit down with us, and this 9 has been enormously helpful. 10 MR. WYLIE: 11 REP. SCHIFF: 12 Sure. I really appreciate you doing it. 13 MR. WYLIE: Well, it's my pleasure. It's why 14 I came. 15 start looking at it, and if there's information that's 16 helpful, I want to provide it. I want as many authorities or legislators to 17 REP. SCHIFF: 18 MR. WYLIE: 19 (A brief recess was taken.) 20 21 22 Thank you. Sure. I'm really looking forward to connecting with your folks in the U.K. -MR. WYLIE: Sure. Page 169 1 2 -- that are doing the investigation. 3 MR. WYLIE: 4 5 I appreciate how long you've been here today, and -- 6 MR. WYLIE: 7 8 Sure. Oh, no worries. I know you've had a long couple of days here, but we just have a few questions. 9 MR. WYLIE: 10 Yeah, sure. And then you can go on your way. 11 MR. WYLIE: 12 Sure. I'm curious if any of your other 13 colleagues from SCL or Cambridge Analytica have also 14 come forward as a whistleblower as you have, other than 15 Brittany Kaiser, I believe. 16 MR. WYLIE: There are numerous people who have 17 provided either myself or journalists or the 18 parliamentary inquiry with information, documents, or 19 testimony. 20 record. Most of those people have done it off the 21 Even after I've come out, they still are 22 extremely wary of this company, but there are many Page 170 1 people who have information and have provided it either 2 to myself or to others. 3 Just a point on Brittney Kaiser. I take 4 slight issue with the term "whistleblower." 5 just my personal perspective. 6 company until very recently and only came forward after 7 it was apparent that that relationship was no longer 8 fruitful. 9 This is She worked at the When you actually look carefully at her 10 testimony to the British Parliament, she actually is 11 very particular in only revealing information about 12 what other people have done and, in particular, only 13 revealing information or focusing that information not 14 only what other people have done, but people outside of 15 the company. 16 The other piece of information that I would 17 provide to you, just to add context to Brittney 18 Kaiser's testimony, is that I was told, and I actually 19 have this screen "shotted" from the investigation's 20 editor of Chanel 4 News, so who commissioned the 21 undercover. 22 They interviewed Brittney Kaiser right after Page 171 1 her testimony, and they saw that Alexander Nix was 2 texting her, and one of those texts was at the end of 3 her testimony and said, "Good job. 4 Well done." So just to add some context to that, the other 5 thing I would also add context to is that her sudden 6 switch from working for Cambridge Analytica to becoming 7 a data privacy advocate in the span of a week I find 8 interesting, and the other -- 9 10 11 12 Can I ask: MR. WYLIE: No. did you hire her? She came on -- it was Alexander Nix who decided to bring her on. I mean, she used to work in Libya. A lot of 13 her connections are Libyans, and she's done all kinds 14 of interesting projects around the world, which I'll 15 let you look at. 16 One of the other things that I would just 17 point out also is that these various cryptocurrency 18 companies that she's now promoting actually will 19 benefit from the legislative -- financially benefit 20 from the legislative changes that she's proposing. 21 she is proposing changes to Facebook and then also 22 privacy policies in the United States that would So Page 172 1 actually create a data market that her cryptocurrency 2 companies would financially profit from. 3 So I just -- sorry. I don't mean to go on a 4 long rant about it. 5 point out is that she hasn't actually been cooperative 6 with the British authorities who have consistently 7 requested information, and she continues to not provide 8 information to British authorities. 9 I just -- the other thing that I'd She took an opportunity to go to the 10 parliamentary inquiry because it was televised, and 11 less than 24 hours later, she launched her new 12 cryptocurrency. 13 flew back to Manhattan, and then launched her 14 cryptocurrency the next day. 15 She flew to London, did the testimony, So just -- I'm just providing you with some 16 context as to, you know, Brittney Kaiser because she 17 continuously presents herself as a whistleblower, and I 18 really take issue. 19 The process that I went through was months and 20 months and months of work, including first going to the 21 authorities, making sure they had plenty of time to 22 investigate the matter; doing a lot of the boring, Page 173 1 long, technical stuff that involves reporting stuff to 2 the authorities, working with regulators that she 3 refuses to do. 4 5 And she just -- anyway, sorry. stop myself. 6 7 The work you've done with Channel 4 and that whole undercover story is phenomenal. 8 9 10 I am going to MR. WYLIE: whole thing. Yeah. I helped them set up the So even coming up with that issue, Sri Lanka, because -- 11 It was brilliant. 12 MR. WYLIE: -- because they -- we had to pick 13 a country, and so they have eight offices in India or 14 rather eight sort of partner offices in India. 15 do work in India. 16 So they But Sri Lanka, the politics of Sri Lanka are 17 so convoluted that it's actually really hard to like -- 18 it's convoluted enough that a quick Google search 19 before a meeting, you can't parse out is this a set-up 20 or not, but it's close enough to their Indian partners 21 that they could actually deliver the work relatively 22 easily. Page 174 1 So we eventually even figured out what company 2 and how to approach it and all of that. 3 bizarre. 4 and yeah. Like it was We had like these strategy planning meetings, Anyway, sorry. 5 Shifting gears just a little bit 6 MR. WYLIE: 7 Yeah, yeah. -- a Columbia University 8 technology expert -- 9 MR. WYLIE: 10 Yeah. -- last fall wrote an article on 11 the Website Medium about an intern at Cambridge 12 Analytica who reportedly left source code on GitHub. 13 MR. WYLIE: 14 15 16 Un-huh. Are you familiar with that article? MR. WYLIE: I am, yeah, and I'm not sure if 17 you've seen that that's happened several times. 18 UpGuard, which is a data security firm, also revealed 19 several misconfigurations on GitLab where some of this 20 happened, I think, two or three weeks ago, where some 21 of the source code for Ripon, which was one of the 22 technology platforms that was developed to deploy some So Page 175 1 2 of the data, was also revealed in GitLab. Sorry. I just -- 3 4 The intern who reportedly -MR. WYLIE: 5 6 7 -- this was a gentleman named Michael Phillip. MR. WYLIE: 8 9 Yeah. Yeah. Did you happen to know him? MR. WYLIE: I didn't, no. I'm not sure if he 10 was there when I was there or not, but I didn't 11 actually engage with the interns that much. 12 didn't have a relationship with him. 13 14 15 So, no, I And were you familiar with the data that was left on GitHub? MR. WYLIE: I'm not familiar. I know that 16 there was some kind of information that was left, but I 17 don't recall which. 18 19 20 21 22 Okay. I was just curious if you knew him and knew about that data. MR. WYLIE: Yeah. There were some questions that arose about whether that was possibly left purposefully Page 176 1 for others to access or if potentially Russians could 2 have 3 MR. WYLIE: Yes. Something to -- one of the 4 things that I remember being told about was -- and it 5 never occurred to me that you would do -- and this is 6 why I remember it, because I was just like, "Oh, that's 7 quite clever" -- was that if you do -- so, for example, 8 if you have a PAC that's supporting a particular or 9 that's not -- because a PAC can't technically support a 10 11 candidate. No, sometimes they can. Anyway, if they want footage of a candidate 12 but can't film it directly, that you post it on like 13 YouTube, for example, and call it something very 14 random, and then they just happen to find it, and then 15 they can use it. 16 And so I remember something to the effect of 17 like not with respect to data, but with respect to like 18 content and various things to avoid actually 19 transferring something, to just sort of leave it 20 somewhere and then somebody can find it. 21 22 With respect to leaving data around for people to find, that wasn't something that I was aware of, and Page 177 1 frankly, if I knew about it, I would have stopped it 2 because at least in U.K. that's potentially unlawful 3 under the Data Protection Act. 4 Can I ask on this point? Because 5 I know in Parliament I think you had said that, and you 6 were very clear on this, you don't have any evidence of 7 actual collusion between Mr. Nix or Kogan and Russian 8 interfering in the election. 9 MR. WYLIE: 10 Yeah. But your concern is that because 11 of perhaps insecurity, things aren't protected, Mr. 12 Kogan is going back and forth; Dr. Kogan is going back 13 and forth. 14 MR. WYLIE: 15 Yeah. The Russians maybe knew what he 16 was doing and maybe sought opportunity to do something 17 about it. 18 19 MR. WYLIE: Yeah. Just sort of baseline level, how 20 technically savvy does one have to be? 21 know, hacked or steal the algorithms that you were 22 talking about that CA developed. Say you, you Page 178 1 MR. WYLIE: Un-huh. 2 How technically savvy do you have 3 to be to actually deploy those if your intention was to 4 interfere or weaponize it in some way, you know, 5 baseline level? 6 MR. WYLIE: 7 Sure. You know, perhaps a nation-state 8 actor has these capabilities, but I'm just curious. 9 Like what do you if you come across this data? 10 MR. WYLIE: 11 12 13 14 And, you know, acquire it illegally, putting aside like leaving it in the open. MR. WYLIE: I hear you on that. MR. WYLIE: 17 18 So I think there's a difference between the data and the algorithm. 15 16 Yeah. And also the actual acquisition. Sure. MR. WYLIE: Because the security practices -- 19 let me give you an example of how poor the security 20 practices were. 21 22 That would be very helpful. MR. WYLIE: I got emails, unencrypted emails, Page 179 1 all of the access credentials for the shared servers 2 for the entire company. 3 send you, so it's not this exact password, but it was 4 something to the effect of SCL -- 5 The passwords, which I can One, two, three, four. 6 MR. WYLIE: 7 Brute force attack, it would have taken -- -- 2014. 8 which you don't need to be -- to do a brute force 9 attack, it's like, I mean, you don't have to be very 10 technical. 11 that's not a complicated way of hacking. You have to be slightly technical, but 12 Sure. 13 MR. WYLIE: 14 It's like baby hacking. Sure. 15 MR. WYLIE: If you have just a combination of 16 letters and numbers and it's that long and also even a 17 dictionary attack, you'd start with dictionary attack. 18 The first thing you do is start with like common names 19 and the year or 123. 20 of minutes, and you would have then had access to the 21 servers. 22 That probably would take a couple Right. Page 180 1 2 MR. WYLIE: unencrypted, just all of it just right there. 3 4 Just sort of basically lax of security. 5 MR. WYLIE: 6 7 8 Yes. The access possibilities are there. MR. WYLIE: 9 10 And also that was emailed to me, Atrocious. Okay. MR. WYLIE: That's step one. So first, in terms of the security 11 practices were so poor that you wouldn't even have to 12 be a very good hacker to do it. 13 Google how to hack something and probably figure it out 14 if you kind of know that code. 15 So that's the first point. 16 17 You probably could Right. MR. WYLIE: The second point is then there's 18 two things that you could possibly acquire. 19 acquire data. 20 the electoral register, consumer data or Facebook data. So that would be either, for example, 21 22 You could Right. MR. WYLIE: If you just simply acquired the Page 181 1 Facebook data and you acquired the electoral register 2 and you acquired commercial data, you could just use 3 that to create a custom universe in Facebook if you 4 wanted to. 5 6 7 Right. MR. WYLIE: Like that wouldn't be hard to do. You would just like make a list, and you could do it 8 in a very rudimentary way. 9 men over 40 who are registered as a Republican. 10 13 14 Upload this custom list, right? 11 12 I want to target like white Right. MR. WYLIE: Like so you wouldn't need to be very technically savvy. On the other hand, if you were to pull 15 algorithms, that's where you're getting into like you 16 need to know -- that's where it would be more 17 complicated because you have to actually understand 18 like machine learning and what it is that you're 19 looking at. 20 21 22 Right. MR. WYLIE: And how things -- all the inputs and the rationale of all the code. So that would be Page 182 1 more technically complex. 2 But at the same time, if you just get somebody 3 who -- I mean, it's not hard to find somebody who like 4 understands statistics. 5 Right. 6 MR. WYLIE: 7 You know, and code. Right. 8 MR. WYLIE: You know, data finds, you know, 9 and so my concern is that because the data was, you 10 know, not that secure and that even if it wasn't an 11 offensive hack, if a hack happened, you know, you could 12 have key logged Kogan's computer in Russia. 13 you initially need a USB, -- 14 You know, Sure, sure. 15 MR. WYLIE: -- stick it in, leave it for 30 16 seconds, and then pull it out, and you key logged it, 17 right? 18 19 Understood. MR. WYLIE: So in terms of deploying the 20 algorithms, that would be more complicated, but like, 21 again, imagine you have access to 87 million records, 22 the bulk of which are American citizens, and you're Page 183 1 Russia or any kind of foreign user. 2 that asset and you also have the intention of misusing 3 it -- 4 5 Opportunity, intention, and capability. 6 7 MR. WYLIE: 10 You know, you have ready Right. MR. WYLIE: And also I don't know. Are you familiar with the Garrison Law doctrine? 11 12 Yeah. access to. 8 9 Like if you have Yes. MR. WYLIE: Right. So when you actually look 13 at how Russian military policy, and this is partly why 14 DARPA and DSTL and various military research agencies 15 were shifting gears into non-kinetic warfare or 16 asymmetric warfare, you know, not -- because you've got 17 kinetic warfare is blowing stuff up, and then non- 18 kinetic, which is like cyber information. 19 Because Russia and several other countries and 20 also ISIS and various other -- because the thing about 21 non-kinetic warfare is that it's so much cheaper, and 22 the dirty secret about Russia is that it's super poor. Page 184 1 Their GDP is like Australia, but they have a -- 2 Right. 3 4 MR. WYLIE: -- nuclear arsenal that they can't afford to maintain. 5 Right. 6 MR. WYLIE: And so if you look at the expense 7 that it is, you know, how expensive it is to, you know, 8 buy a series of missiles or a nuclear weapon, right, 9 and then, you know, Eastern European data scientists 10 aren't that expensive. 11 Right. That all makes sense, and 12 I get that just goes to sort of to decide whether 13 Cambridge was willingly trying to help someone but just 14 by what you already know, already the opportunity was 15 there. That's something someone could have maliciously 16 MR. WYLIE: 17 18 -- done something and done something with it. 19 20 Yeah. I think I've only got one other question actually. 21 Go ahead. 22 And hopefully it's quick. Talked Page 185 1 2 on that. Oh, so I don't know if you saw that, but 3 Alexander Nix spoke at a Lisbon Web Summit late last 4 year, and the Wall Street Journal reported that Nix 5 apparently said that he found it, frankly, absurd that 6 Russia interfered in the election. 7 Any insight as to why Mr. Nix would have that 8 specific an opinion that the fact that Russians 9 interfering in the U.S. election is an absurd concept? 10 11 12 I realize that you can't get into his head, of course, but it just seems -MR. WYLIE: 13 Okay. It caught our attention. So 14 inasmuch as you have a thought on it, I think that's 15 just curious. 16 17 MR. WYLIE: probably shouldn't spec -- 18 19 I don't have any particular -- I Right. MR. WYLIE: I shouldn't speculate on his 20 motives for saying that other than just saying that 21 that is the line of his particular clientele, and so it 22 makes sense that he would say that. Page 186 1 Because if he says, "Yeah, Russians interfered 2 with my and helped my client, who I still want to work 3 with," that probably would be -- 4 Thanks. 5 Mr. Wylie, you may know that Dr. 6 Kogan testified yesterday. 7 MR. WYLIE: Yes. I haven't actually seen his 8 testimony yet, although I had the tidbits of 9 highlights, yes. 10 As you might imagine, he disputes 11 some of your claims. 12 MR. WYLIE: 13 14 Un-huh. Did you set up a company when you left SCL? 15 MR. WYLIE: Not when I left. When I was at 16 Cambridge Analytica, I set up a company because the 17 majority of people -- while Cambridge Analytica was 18 being set up, the majority of people were engaged as 19 consultants rather than employees because, again, they 20 were still working out how to actually set up the 21 company. 22 And also, in the U.K., I don't know that it's Page 187 1 the thing here, but in the U.K., you have to pay 2 additional taxes per employee. 3 4 So it was cheaper. So I wasn't the only person at CIA who had a consulting company. 5 Sure. When you left SCL, did you 6 engage in any type of transaction with Dr. Kogan, any 7 type of data sharing arrangement? 8 9 10 11 MR. WYLIE: Yeah. been super clear on this. So I had access, and I've I had access to data or, rather, I had access to a portion of the data. I also agreed with Kogan that I would also 12 help him acquire new data sets, and I'll be super clear 13 on this, to help his institute. 14 this well documented in writing, that he was going to 15 be setting up an academic institute. He told me, and I have 16 I thought that this was a very good idea, and 17 several professors at Cambridge were very excited about 18 the notion of having a very large data set where you 19 could explore social construct in a quantitative way, 20 and that it would have been sort of the first sort of 21 test case for a new field of or an emerging field of 22 research in computational sociology or computational Page 188 1 psychology. 2 He did not tell me, and I then found out, that 3 he went back after I had left to meet with Alexander 4 Nix where he proposed a commercial contract, and he was 5 going to -- because he told me that he was intending on 6 finalizing or completing his relationship with 7 Cambridge Analytica, and that he was then going to set 8 up this institute, which I was actually very excited 9 about. 10 And when I found out that he went back and 11 didn't tell me about it to Alexander Nix and pitched a 12 commercial project with the data, I told him to go 13 away. 14 15 That's actually very helpful. That helps me fill in the picture. 16 MR. WYLIE: And can I just be super clear on 17 something? 18 any way with this Facebook data set who has never sold 19 the data. 20 in using that data or derivatives of that data for 21 client projects. 22 selling that data or derivatives of that data. I am the only person that was involved in Okay? Cambridge Analytica has been involved Dr. Kogan has been involved in Page 189 1 I am the only person who had access to that 2 data who has never sold it, just to be super clear on 3 that. 4 Analytica has how set up a Website on how I'm not a 5 whistleblower. Because there is this narrative that Cambridge 6 7 You know, I believe you. MR. WYLIE: 8 9 10 And I just want to give you an opportunity. MR. WYLIE: 11 12 Yeah. Oh, no, no, no, that's fine. To resolve that. MR. WYLIE: Yeah, that's fine. It's just I 13 find it very -- the thing that I find frustrating is 14 that like I have gone out of my way to accept my share 15 of responsibility in this project. 16 access to the data. 17 it, which is why I'm talking about it 18 played a fundamental role in facilitating that project 19 and, more generally, setting up Cambridge Analytica. 20 I don't deny I had I played a -- and I deeply regret -- you know, And, you know, I'm disappointed that, you 21 know, Aleksandr Kogan tried to obfuscate his 22 responsibility of Facebook, tried to obfuscate the Page 190 1 responsibility of Cambridge Analytica. 2 starting to just blame everyone else, and it would just 3 be so much simpler if he just said, "Yes, I did 4 something stupid. 5 frustrating. I screwed up." 6 7 8 And I just find it Well, hopefully some of these committees will get to the bottom of these activities. MR. WYLIE: 9 10 Everyone is Yeah. But again, on behalf of the Ranking Member, we appreciate your patience. 11 And Speaker Pelosi for -- 12 And Speaker Pelosi. 13 14 So that completes -MR. WYLIE: Because I felt really cool to like 15 meet Nancy Pelosi, and she's just sitting there. 16 was, "oh, my God, you're great. 17 (Laughter.) 18 19 20 21 22 It Love you." We can go off the record. (Whereupon, at 7:37 p.m., the interview was completed.) * * * * *