No. 19-783

IN THE
___________________

NATHAN VAN BUREN,
Petitioner,
v.
UNITED STATES,
Respondent.
___________________

ON WRIT OF CERTIORARI TO THE UNITED STATES
COURT OF APPEALS FOR THE ELEVENTH CIRCUIT
___________________

BRIEF OF AMICUS CURIAE THE MARKUP IN
SUPPORT OF PETITIONER
__________________

Nabiha Syed
THE MARKUP

900 Broadway, Suite 202
New York, NY 10159
(347)894-3746
nabiha@themarkup.org

Katherine M. Bolger*
John M. Browning

DAVIS WRIGHT TREMAINE LLP
1251 Avenue of the Americas
New York, NY 10020
(212) 489-8230
katebolger@dwt.com

David M. Gossett

DAVIS WRIGHT TREMAINE LLP
1919 Pennsylvania Ave. NW
Washington, DC 20006

*Counsel of Record
Counsel for Amicus Curiae

 i

TABLE OF CONTENTS
Page
TABLE OF AUTHORITIES .................................... iii
INTEREST OF AMICUS CURIAE ......................... 1
INTRODUCTION AND SUMMARY OF
ARGUMENT ....................................................... 2
ARGUMENT ............................................................ 6
I.

DATA JOURNALISM IS A VITAL
FORM OF JOURNALISM. ......................... 7

II.

THE MARKUP USES DATA
JOURNALISM TECHNIQUES TO
REPORT ON THE TECHNOLOGY
SHAPING OUR SOCIETY. ...................... 11
A. The Markup Gathers News by
Observing What Can Be Done on
Online Platforms............................... 13
B.

Data Scraping Is a Vital Tool for
Data Journalists. .............................. 15

III. THE FIRST AMENDMENT
PROTECTS THE ONLINE
NEWSGATHERING PRACTICES OF
DATA JOURNALISTS. ............................ 21
A. Routine Newsgathering Activities
Cannot Be Criminalized Under the
First Amendment Just Because
They Were Conducted Online. ......... 21

 ii

B.

The CFAA Should Be Interpreted
to Avoid Violating the First
Amendment ....................................... 27

CONCLUSION ....................................................... 29

 iii

TABLE OF AUTHORITIES
Page(s)
Cases
Bartnicki v. Vopper,
532 U.S. 514 (2001) .............................................. 24
Bond v. United States,
572 U.S. 844 (2014) .............................................. 27
Cox Broadcasting Corp. v. Cohn,
420 U.S. 469 (1975) .............................................. 22
Florida Star v. B.J.F.,
491 U.S. 524 (1989) ........................................ 23, 25
Garrison v. Louisiana,
379 U.S. 64 (1964) ................................................ 22
HiQLabs, Inc. v. LinkedIn Corp.,
938 F.3d 985 (9th Cir. 2019) ................................ 25
J.H. Desnick v. American Broadcasting Cos.,
44 F.3d 1345 (7th Cir. 1995) ............................... 14
Oklahoma Publ’g Co. v. District Ct.,
430 U.S. 308 (1977) .......................................... 6, 22
Packingham v. North Carolina,
137 S. Ct. 1730 (2017) ................................ 6, 22, 26
Reno v. ACLU,
521 U.S. 844 (1997) .......................................... 6, 26

 iv

Sandvig v. Barr,
--- F. Supp. 3d ---, 2020 WL 1494065
(D.D.C. Mar. 27, 2020), appeal filed,
No. 20-5153 (D.C. Cir. May 28, 2020) ......... 5, 6, 28
Smith v. Daily Mail Publ’g Co.,
443 U.S. 97 (1979) ...................................... 5, 23, 25
Constitutional Provisions and Statutes
U.S. Const. amend. I ......................................... passim
18 U.S.C. § 1030 ................................................ passim
Other Authorities
Julia Angwin, A Letter from the Editor, THE
MARKUP (Feb. 25, 2020) ....................................... 12
Julia Angwin et al., Despite Disavowals,
Leading Tech Companies Help Extremist
Sites Monetize Hate, PROPUBLICA (Aug. 19,
2017) ....................................................................... 9
Julia Angwin, et al., Facebook Lets Advertisers
Exclude Users by Race, PROPUBLICA (Oct.
28, 2016) ............................................................... 11
Julia Angwin et al., Machine Bias,
PROPUBLICA (May 23, 2016) ............................ 3, 11
Nellie Bly, TEN DAYS IN A MAD HOUSE (1887) .......... 14
Jacquellena Carrero, Note, Access Granted: A
First Amendment Theory of Reform of the
CFAA Access Provision, 120 COLUM. L. REV.
131 (2020) ............................................................... 4

 v

Bill Dedman, The Color of Money, ATLANTA
JOURNAL-CONSTITUTION (1988).................. 3, 7, 8, 9
John Keegan, Blue Feed, Red Feed, WALL
STREET JOURNAL (Aug. 19, 2019)......................... 19
Nathaniel Lash, Why Pinellas County is the
worst place in Florida to be black and go to
public school, TAMPA BAY TIMES (Aug. 12,
2015) ....................................................................... 3
Wesley Lowery et al., 994 people shot dead
by police in 2015, WASH. POST (2015) .................... 3
Rob O’Dell et al., Copy, paste, legislate, USA
TODAY (Apr. 3, 2019) ............................................ 10
Komal S. Patel, Note, Testing the Limits of the
First Amendment: How Online Civil Rights
Testing Is Protected Speech Activity, 118
COLUM. L. REV. 1473 (2018) ................................... 4
Aaron Sankin & Maddy Varner, Want to
Find a Misinformed Public? Facebook’s
Already Done It, THE MARKUP (Feb. 25,
2020) ..................................................................... 13
Ken Schwencke et al., Nonprofit Explorer,
PROPUBLICA .......................................................... 19
Nate Silver et al., Latest Polls,
FIVETHIRTYEIGHT.COM (June 7, 2020)............. 3, 10
Nabiha Syed, President, The Markup, Salant
Lecture on Freedom of the Press at the
Harvard Kennedy Center (Nov. 14, 2018) .......... 17

 vi

Nabiha Syed, A Letter from the President,
THE MARKUP (Feb. 25, 2020) ............................... 12
Ida B. Wells-Barnett,
LYNCHING AND THE EXCUSES FOR IT (1901) ........... 2
Ida B. Wells-Barnett,
THE RED RECORD (1895)......................................... 2
Leon Yin & Adrianne Jeffries, Show Your
Work: How We Examined Gmail’s
Treatment of Political Emails, THE
MARKUP (Feb. 26, 2020) ....................................... 15
Sandhya Somashekhar et al., Black and
Unarmed, WASH. POST (Aug. 8, 2015) ................... 9
Taken, PULITZER CENTER .......................................... 10
Wayback Machine, INTERNET ARCHIVE .................... 16

 1
INTEREST OF AMICUS CURIAE1
The Markup is a nonprofit news organization that
conducts data-driven investigations into how
powerful institutions use digital technology to
reshape our society. Launched in February 2020, The
Markup has assembled a staff of journalists who
specialize in data journalism – also known as
computational journalism – and publishes their work
online at themarkup.org. The core mission of The
Markup is to use the tools of data journalism to
investigate how technology works, particularly when
it is wielded by the online platforms, government
entities, and other organizations that play an
increasingly prominent role in our online (and real
world) lives. The Markup thus has a compelling
interest in ensuring that the newsgathering practices
that make data journalism a valuable tool for holding
the powerful accountable in our modern democracy
are not chilled or made illegal by overbroad
application of data security laws.

1 All parties have provided their consent to

the filing of this brief.
No counsel for a party authored this brief in whole or in part, and
no person or entity other than amicus and its counsel made a
monetary contribution to the preparation or submission of this
brief.

 2
INTRODUCTION AND
SUMMARY OF ARGUMENT
In 1895, pioneering journalist Ida B. Wells
published The Red Record: Tabulated Statistics and
Alleged Causes of Lynching in the United States, 18921894. Ida B. Wells-Barnett, THE RED RECORD (1895).
In that publication and several others like it, Wells
combed through public records and compiled
statistical charts to demonstrate the frequency with
which men, and particularly Black men, were
lynched. The purpose of Wells’ statistical analysis was
not merely actuarial. Rather, it served as the raw
material for her news reporting, which changed the
way lynching was perceived and debated, both by its
critics and its supporters. The power of Ms. Wells’
reporting lay in her devotion to facts. As she wrote,
“[n]o good result can come from any investigation
which refuses to consider the facts. A conclusion that
is based upon a presumption, instead of the best
evidence, is unworthy of a moment’s consideration.”
Ida B. Wells-Barnett, LYNCHING AND THE EXCUSES FOR
IT (1901). By carefully gathering and marshalling
indisputable facts from public records scattered across
the nation, Ms. Wells revealed the frightful scale of
lynching in this country; without her work, the truth
may have remained hidden in plain sight.
The Markup continues the tradition of data
journalism that Ms. Wells pioneered, the need for
which could not be greater today. Like Ms. Wells, The
Markup seeks to inform and influence public debate
by marshaling the “best evidence” available – which is
now increasingly online. Whereas Ms. Wells forced
society to confront the evils of lynching by manually

 3
searching paper records, modern data journalists
analyze enormous datasets to identify abuses of power
in the modern world. With increasingly powerful and
sophisticated tools at their disposal, data journalists
have exposed countless injustices, including: systemic
racial discrimination by banks refusing to issue loans
to Black customers; the re-segregation of Florida
schools after integration was no longer enforced; and
flaws in algorithms designed to assist with
sentencing, which incorrectly rated Black defendants
as posing higher risks of reoffending than White
defendants.2 Data journalists have also created public
databases that track police shootings and offer
sophisticated insight into our political system by
aggregating individual polls.3
All of these projects illuminate the invisible forces
at work in peoples’ lives. The common thread running
back to Ms. Wells is that journalists can use data to
observe and critique powerful entities or entrenched
systems. It is impossible to report on systematic
Bill Dedman, The Color of Money, ATLANTA JOURNALCONSTITUTION (1988), http://powerreporting.com/color/color_of_
money.pdf; Nathaniel Lash, Why Pinellas County is the worst
place in Florida to be black and go to public school, TAMPA BAY
TIMES (Aug. 12, 2015), https://www.tampabay.com/news/
education/k12/why-pinellas-county-is-the-worst-place-in-floridato-be-black-and-go-to/2241065/; Julia Angwin et al., Machine
Bias, PROPUBLICA (May 23, 2016), https://www.propublica.org/
article/machine-bias-risk-assessments-in-criminal-sentencing.
2

Wesley Lowery et al., 994 people shot dead by police in 2015,
WASH. POST (2015), https://www.washingtonpost.com/graphics/
national/police-shootings/; Nate Silver et al., Latest Polls,
FIVETHIRTYEIGHT.COM (June 7, 2020), https://projects.fivethirty
eight.com/polls/president-general/national/.
3

 4
trends in our society caused by powerful
governmental or private actors – such as racial
discrimination, police misconduct, or anti-competitive
behavior – without collecting enough information to
understand the big picture. Indeed, many of these
abuses would remain invisible if journalists lacked
the power to collect raw data from online sources.
Data journalism thus enables citizens to comprehend
our increasingly complex world and, in doing so, gives
us all the power to demand change.
Yet the newsgathering techniques that make data
journalism possible could be effectively outlawed by
the broad reading of the CFAA that the government
urges and that the Eleventh Circuit endorsed. As the
Eleventh Circuit acknowledges, its interpretation of
the CFAA empowers the operator of any website “to
legislate what counts as criminal behavior through
their internal policies or terms of use.” Pet. App. 27a.
In practice, this means that journalists may risk
criminal liability – and prison time – for violating
even the most trivial terms of service for any website
they investigate. Worse still, any website owner
seeking to avoid legitimate oversight could instantly,
or perhaps even retroactively, make it a crime for
journalists to investigate their services by
deliberately tweaking the terms of service to prohibit
newsgathering activities. Many websites have already
taken such steps. Jacquellena Carrero, Note, Access
Granted: A First Amendment Theory of Reform of the
CFAA Access Provision, 120 COLUM. L. REV. 131, 134
(2020); Komal S. Patel, Note, Testing the Limits of the
First Amendment: How Online Civil Rights Testing Is
Protected Speech Activity, 118 COLUM. L. REV. 1473,
1494 (2018).

 5
The chilling effect on data journalism is obvious.
The Markup’s reporters and other data journalists
routinely harvest data from government databases
and digital platforms. This is not “hacking.” Rather,
The Markup’s reporters gain authorized access to the
data – often by setting up legitimate accounts – and
have no more (or less) access than any other
authorized user. They do not intentionally circumvent
technology designed to limit access to non-public
information and they endeavor – as far as possible –
to abide by the terms of service for every website they
investigate. And, because its reporters have nothing
to hide, The Markup publishes “Show Your Work”
articles describing how data was collected for any
given news article. But even with these efforts, The
Markup’s journalists could violate the CFAA because
it is impossible to guarantee perfect compliance with
every website’s “protean” terms of use, which are
invariably “long, dense, and subject to change.”
Sandvig v. Barr, --- F. Supp. 3d ---, 2020 WL 1494065,
at *10 (D.D.C. Mar. 27, 2020), appeal filed, No. 205153 (D.C. Cir. May 28, 2020).
The First Amendment cannot tolerate such
arbitrary restrictions on newsgathering. In an
unbroken line of cases establishing the right to
publish lawfully obtained information concerning a
matter of public interest, this Court has made it clear
that the government cannot criminalize “routine
newspaper reporting techniques.” Smith v. Daily Mail
Publ’g Co., 443 U.S. 97, 103 (1979). This Court has
also held that journalists cannot be punished for
obtaining and publishing newsworthy information
after it has been “publicly revealed,” even if that
information was legally required to remain secret.

 6
Oklahoma Publ’g Co. v. District Ct., 430 U.S. 308, 311
(1977). It makes no difference to the First Amendment
whether the information was gathered online or in an
analog format. As this Court has recognized, the
Internet is a “vast democratic forum[],” Reno v. ACLU,
521 U.S. 844, 868-69 (1997) – the “modern public
square.” Packingham v. North Carolina, 137 S. Ct.
1730, 1737 (2017). Data journalism serves an
important role in helping Americans observe and
understand the forces at play in that public square
that might otherwise remain hidden. A statute that
allows powerful forces like the government or wealthy
corporate
actors
to
unilaterally
criminalize
newsgathering activities by blocking these efforts
through the terms of service for their websites would
violate the First Amendment.
But this Court can avoid a collision course with the
First Amendment by giving the CFAA the rational
interpretation urged by petitioner. As Judge Bates
recently explained in rejecting the government’s
broad reading of the CFAA, “[i]f this Court determines
that the [CFAA] does not actually criminalize”
journalists for “violating consumer websites’ terms of
services,” then “the Court need not dive … into the
First Amendment policy.” Sandvig, 2020 WL
1494065, at *7. Accordingly, this Court should reverse
the Eleventh Circuit and hold that, if a person has
authority to access material on a computer system,
that access does not violate the CFAA – even if it is for
a reason beyond those for which access was granted.
ARGUMENT
The Markup is at the vanguard of data journalism,
which has a venerable tradition but has rapidly

 7
evolved in response to the explosion of data now
available to reporters. Essentially, data journalists
use digital tools to do what investigative reporters
have always done: gather the information necessary
to reveal matters of public concern that would
otherwise remain hidden, and then translate it into
news stories for the public’s benefit. The First
Amendment protects basic newsgathering techniques
that make journalism possible. This Court should
reject a broad reading of the CFAA that would thwart
those First Amendment principles by potentially
outlawing the tools that The Markup and other news
organizations use to gather data online for the
purpose of investigative reporting.
I.

Data Journalism Is a Vital Form of
Journalism.

The roots of data journalism can be found in the
public-service tradition of investigative journalism.
The tradition stretches back at least as far as Ms.
Wells and other reporters who manually gathered,
analyzed, and published data. This form of journalism
is necessary in order to identify structural problems
or patterns in our society that cannot be discerned by
looking at incidents in isolation. For instance,
Ms. Wells’ masterful use of data disproved the fiction
that lynching was sporadic and showed it for what it
was – a systemic campaign of terror.
Another classic example of data journalism is The
Color of Money, a Pulitzer-Prize winning series of
reports about racially discriminatory mortgage
lending practices that appeared in The Atlanta
Journal-Constitution in 1988. Dedman, supra note 2.
For his classic report, Bill Dedman used the Freedom

 8
of Information Act to obtain “computer tape”
containing lenders’ reports relating to a total of
109,000 loan decisions made in metro Atlanta
between 1981 and 1986. Id. Using statistical analysis
and publicly available demographic information, Mr.
Dedman was able to determine that “[w]hites receive
five times as many home loans from Atlanta’s banks
and savings and loans as blacks of the same income –
and that gap has been widening each year.” Id. at 3.
The power of Mr. Dedman’s reporting lay once again
in his use of data, which proved that the
discrimination was not intermittent but rather an
institutional practice across a whole city. Once these
systemic injustices were revealed, it was possible to
effect meaningful change.
In the years since, the growth of personal
computing and the Internet has caused an explosion
in the amount of data that our society generates –
which, in turn, has expanded the horizon of what is
possible through data journalism. Since the
beginning, the unique strength of data journalism has
been its ability to discern and articulate structural
problems with our democracy by identifying trends
that are simply not apparent from focusing on one-off
anecdotes. A prime example of the insights made
possible by data journalism is The Washington Post’s
initiative led by Wesley Lowery to create a national
database tracking police shootings, which won the
2015 Pulitzer Prize. The information in the database
was used as the basis for a series of reports about the
frequency of police shootings and who the victims
were likely to be; it exposed the reality that “unarmed
black men are seven times more likely than whites to
die by police gunfire.” Sandhya Somashekhar et al.,

 9
Black and Unarmed, WASH. POST (Aug. 8, 2015),
https://www.washingtonpost.com/sf/national/2015/08/
08/black-and-unarmed/?itid=sf_. There is a clear link
between this reporting and Wells’ earlier work on
lynching. In both cases – and Dedman’s Color of
Money series – the data allowed journalists to observe
and expose injustices that would have otherwise
remained hidden.
Indeed, within the last decade, news organizations
like The Markup, ProPublica, The Intercept, Five
Thirty Eight, and others – as well as journalists at
traditional news organizations – have pioneered new
analytical reporting techniques that have made it
possible to use data to reveal hitherto unseen truths
about the way our world works. The discipline has
developed rapidly to adapt to a society that is
increasingly awash with data, and data journalists
have won at least 3 Pulitzer Prizes as well as
countless other awards. The vibrant field of data
journalism also covers a wide range of subject matter
and generates a diverse range of output, from hardhitting investigative reports about matters of intense
public concern to innovative graphical reporting that
uses pictures or video to bring data to life. A few
examples include:


ProPublica reporters demonstrated how
certain online hate groups were able to
monetize their content despite websites’ claims
to the contrary. Julia Angwin et al., Despite
Disavowals, Leading Tech Companies Help
Extremist Sites Monetize Hate, PROPUBLICA
(Aug. 19, 2017),
https://www.propublica.org/

 10
article/leading-tech-companies-help-extremistsites-monetize-hate.


Taken is a collaborative reporting effort to
collect, analyze and publish data-driven
investigative reports on the prevalence of civil
asset forfeiture across the nation, which is a
process by which police departments seize
property from civilians and sell or keep it for
their own benefit. Taken, PULITZER CENTER,
https://taken.pulitzercenter.org/.



USA Today exposed the prevalence of “fill-inthe-blank” legislation written by corporations,
interest groups or lobbyists, revealing that
legislatures introduced more than 10,000
model bills within the last eight years. Rob
O’Dell et al., Copy, paste, legislate, USA TODAY
(Apr. 3, 2019),
https://www.usatoday.com/
pages/interactives/asbestos-sharia-law-modelbills-lobbyists-special-interests-influencestate-laws/



Five Thirty Eight and others aggregate data
from different pollsters to create “polls of polls”
that more accurately reflect public opinion. See
Nate
Silver
et
al.,
Latest
Polls,
FIVETHIRTYEIGHT.COM (June 7, 2020), https://
projects.fivethirtyeight.com/polls/presidentgeneral/national.

Rather than speaking solely about data journalism
in the abstract, however, we think it will help the
Court to understand the implications of this case to
explain how The Markup operates – and thus how the
government’s interpretation of the CFAA could

 11
effectively prohibit The Markup from continuing to
function.
II.

The Markup Uses Data Journalism
Techniques to Report on the Technology
Shaping Our Society.

The Markup takes a scientific approach to news
reporting that was created by the editor-in-chief and
founder of The Markup, Julia Angwin. Ms. Angwin
previously led investigative teams at ProPublica and
The Wall Street Journal. She is a winner and two-time
finalist for the Pulitzer Prize in journalism. The
Markup employs a roster of data journalists, who
have, over the course of their careers, used datadriven journalism to reveal racial biases in criminal
sentencing software that discriminates against black
defendants; document housing discrimination on
Facebook; and break other news stories that would
not have been possible without access to data and
sound analytical reporting techniques. See, e.g., Julia
Angwin et al., Machine Bias, PROPUBLICA (May 23,
2016) (Pulitzer Prize Finalist for Explanatory
Reporting); Julia Angwin, et al., Facebook Lets
Advertisers Exclude Users by Race, PROPUBLICA (Oct.
28, 2016). These journalists follow the “The Markup
Method” of reporting, which is a three step process:


Build. We ask questions and collect or build
the datasets we need to test our hypotheses.



Bulletproof.
We bulletproof our stories
through a rigorous review process, inviting
external experts and even the subjects of
investigations to challenge our findings.

 12


Show our work. We share our research
methods by publishing our datasets and our
code. And we explain our approach in detailed
methodological write-ups.4

In short, The Markup’s journalists are “dedicated to
deciphering the invisible infrastructure of our world”
because its readers “deserve to understand the forces
that shape … reality.” Nabiha Syed, A Letter from the
President,
THE MARKUP
(Feb.
25,
2020),
https://themarkup.org/2020/02/25/president-letternabiha-syed.
Much like the scientific method, The Markup
Method is about developing conclusions from
observable facts. In addition to being rigorous and
transparent, The Markup – like many, but not all,
other organizations devoted to data journalism –
makes reasonable efforts not to intentionally violate
the terms and conditions of the digital platforms it
investigates. The Markup also aims to be transparent
when it seeks to access data (to the extent possible)
and does not deliberately circumvent technological
measures website operators have imposed to restrict
access to information. In other words, The Markup
does not intentionally “hack” into sources of
information that it is not authorized to see by
breaking security measures and, to the extent
possible, it takes precautions to avoid exceeding the
terms of its authorized access.

See Julia Anguin, A Letter from the Editor, THE MARKUP
(Feb. 25, 2020), https://themarkup.org/2020/02/25/editor-letterjulia-angwin.
4

 13
But even taking these precautions, data
journalists could still face criminal liability for even
minor or inadvertent infractions of a website’s terms
and conditions if this Court adopts the Eleventh
Circuit’s broad reading of the CFAA. That
interpretation potentially criminalizes two of the most
fundamental methods that data journalists like The
Markup use to collect data:
accessing publicly
available digital platforms to investigate how they
work, and using automated tools to extract content
from web sites (also known as data scraping).
A. The Markup Gathers News by Observing
What Can Be Done on Online Platforms.
One of the most elementary ways that The Markup
gathers information is for one of its reporters to
simply sign up for an account on an online service –
whether private-sector or governmental – and use
that account to observe how the website works.
For instance, The Markup recently published an
article that investigated whether Facebook allowed
advertisers on its website to target people that
Facebook identified as being interested in
“pseudoscience,” and thus possibly susceptible to
dangerous conspiracy theories. See Aaron Sankin &
Maddy Varner, Want to Find a Misinformed Public?
Facebook’s Already Done It, THE MARKUP
(Feb. 25, 2020)
https://themarkup.org/coronavirus/
2020/04/23/want-to-find-a-misinformed-publicfacebooks-already-done-it. In order to gather the facts
needed to report this story, Maddy Varner, an
investigative data journalist from The Markup, logged
onto Facebook using her real name and The Markup’s
official account. Id. Once she obtained authorized

 14
access to the platform, she purchased advertisements
and was able to target more than 50 million people on
Facebook and Instagram who were deemed to be
interested in “pseudoscience.” Id. The Markup also
found
evidence
that
users
interested
in
“pseudoscience”
were
being
targeted
with
advertisements for products that were connected to
prominent conspiracy theories – such as hats
purporting to protect the user from electromagnetic
radiation from 5G cellular equipment, which
conspiracy theories have touted as a possible cause of
COVID-19. Facebook ultimately removed its option to
target users interested in “pseudoscience” in response
to the story. Id.
The reporting techniques that The Markup used to
test and report about Facebook’s safeguards against
the sale of pseudoscience items are hardly novel or
controversial. It is a modern twist on a venerable
reporting practice that stretches back at least as far
as Nellie Bly, who posed as a psychiatric patient to
gather information for a series of classic undercover
reports about brutal conditions at the Women’s
Lunatic Asylum on Roosevelt Island. See Nellie Bly,
TEN DAYS IN A MAD HOUSE (1887). See also J.H.
Desnick v. American Broadcasting Cos., 44 F.3d 1345,
52-56 (7th Cir. 1995) (holding that undercover
recordings made by journalists posing as patients of
an ophthalmologist for news report were lawful). And
there is no question that the reporting served a public
goal – once Facebook became aware of the
pseudoscience category, it took steps to eliminate it.
Even so, these routine newsgathering efforts –
which would be entirely unremarkable in the analog

 15
world – could potentially expose The Markup and its
reporters to severe criminal liability under the
Eleventh Circuit’s reading of the CFAA. Instead of
acting responsibly, as they did in this example, all
Facebook needed to do was change their terms of
service to ban newsgathering and The Markup
reporters would face criminal liability. Similarly, the
Eleventh Circuit’s rule would also have put The
Markup’s reporters at the mercy of Google when these
reporters signed up for a Gmail account in order to
test how Google handled emails from political
campaigns. See Leon Yin & Adrianne Jeffries, Show
Your Work: How We Examined Gmail’s Treatment of
Political
Emails,
THE MARKUP
(Feb.
26,
2020). Moreover, website operators seeking to avoid
scrutiny could change their terms of service to
prohibit newsgathering and unilaterally – or perhaps
even retroactively – make it a crime for any journalist
to access a website in order to gather information for
a news story. Allowing the investigated entity to so
unilaterally control the fate of the investigator
overturns long standing journalistic principles and
burdens valuable newsgathering.
B. The
Markup
Uses
Data
Scraping
Technology to Gather Information for its
News Reporting.
Data scraping is the second common practice of
data journalism used by The Markup that is imperiled
by the Eleventh Circuit’s overly broad reading of the
CFAA. “Data scraping” is a term of art that refers to
an automated process to extract content from a
webpage using a specialized software tool or piece of
computer code. It allows journalists to observe vast

 16
quantities of data at scale. A well-known example of
data scraping is the Internet Archive’s Wayback
Machine, which uses software to “crawl” through the
world wide web by systematically visiting webpages
at regular intervals and automatically downloading
copies of any publicly available information. The
result is a database that allows users to go “back in
time” to see what websites looked like in the past,
even after their owners update their content. 5
In the journalism context, reporters routinely use
automated data collection technology to assemble and
review large datasets of publicly available
information. The technology has two crucial benefits.
The first is that it allows journalists to gather and
organize data on a scale that would be practically
impossible to do manually. While it would be
theoretically possible for a reporter, for instance, to
collect the same information manually from every
website, scraping allows journalists to gather
information on a scale that is commensurate with the
gigantic amounts of information available. The second
crucial function of scraping is to allow journalists to
observe and understand trends. In the past, a
journalist might carry out a consumer survey in the
real world or systematically visit public record offices
across the country to pull certain types of records.
See Wayback Machine, INTERNET ARCHIVE, https://archive.
org/web/. Every webpage that is copied by the Internet Archive’s
software (known as a “web crawler”) is placed into an online
database, where it can be searched by the time the webpage was
copied and the URL address from which it was taken. See
Wayback Machine General Information, INTERNET ARCHIVE,
https://help.archive.org/hc/en-us/articles/360004716091Wayback-Machine-General-Information.
5

 17
Scrapers are tools that vastly expand journalists’
ability to observe trends that can serve as the basis
for news reporting, often in real time. As entities
increasingly use data to widen their own audiences
and customer bases, scrapers take advantage of the
same technology to gather information. While it was
possible for Ida B. Wells to gather the information she
needed to expose lynching by physically visiting
public records offices, scrapers make it possible for
reporters to adapt the tradition of data-driven
journalism to the demands of newsgathering in the
digital realm.
As the President of The Markup put it at the 2018
Salant Lecture on Freedom of the Press at Harvard
University, “[w]e need the media’s truth telling role to
be deciphering the hidden systems that result in
observable facts.” Nabiha Syed, President, The
Markup, Salant Lecture on Freedom of the Press at
the Harvard Kennedy Center (Nov. 14, 2018),
https://shorensteincenter.org/nabiha-syed-2018-salan
t-lecture-freedom-press/. Scraping allows journalists
to translate the data that websites make available to
the public into a scale and form that can be observed
critically.
A centerpiece of The Markup’s work is gathering
large amounts of data by automated means, and its
journalists thus routinely design computer code that
collects information automatically or through
systematic processes. The Markup – like most other
data journalists – takes precautions so that its
automated data retrieval applications do not
intentionally access and collect information that
would not available to an authorized human user. The

 18
Markup also endeavors to design its scraping codes to
avoid violating terms of service or circumventing
technological restrictions on access. Finally, in an
effort to maintain transparency and avoid any
appearance of impropriety in the use of automated
data gathering technology, The Markup publishes the
code it uses to analyze data for its investigations and
explains how that code operates in its “Show Your
Work” articles.
A prime example of the unique possibilities of data
scraping is The Markup’s forthcoming “Blacklight”
project. This privacy inspection tool uses software
designed by The Markup to ascertain the ways in
which a website is secretly collecting data about its
visitors. Blacklight works by using automated
techniques to imitate the process of a human visiting
a website. Blacklight launches a software program
that visits a website and probes it to determine what
types of data are being collected. Essentially it is the
mirror image of the usual process whereby a website
takes information to track a user; here, The Markup’s
software examines a website and reports back what
information is being taken. This tool – which The
Markup will make freely available to the public –
shows users how websites track them using their own
personal data. In addition to allowing members of the
public to use the tool, The Markup also intends to
itself use Blacklight as the basis for investigative
reporting on how website operators use surveillance
technology to track users. Without automated
scraping technology, this project would be impossible
and the insights it can offer the public about how they
are being tracked online would never come to light.

 19
A broad reading of the CFAA would imperil this
newsgathering by The Markup and similar
newsgathering by other news organizations that seek
to collect data in bulk using scrapers in order to
inform the public. For example,


Red Feed/Blue Feed, a project of the Wall Street
Journal, which used data scraped from
Facebook to compare the types of articles
served to Facebook users across the political
spectrum. John Keegan, Blue Feed, Red Feed,
WALL STREET JOURNAL (Aug. 19, 2019),
https://graphics.wsj.com/blue-feed-red-feed/.



ProPublica has accumulated a database of up
to 3 million tax returns from tax-exempt
organizations, which it has made available to
the public to search for free. See Ken
Schwencke et al., Nonprofit Explorer,
PROPUBLICA,
https://projects.propublica.org/
nonprofits/.



Public interest organizations have used
scraping to make public documents more freely
accessible to the public, like judicial decisions
on PACER. See, e.g., Free Law Project, https://
free.law/recap/.

In all of these cases, the data being collected by the
scrapers remains available to any authorized user of
the service from which it was obtained, and does not
require journalists to circumvent technological
barriers in order to gain access (i.e., hacking). The
data scraping employed is merely a mechanical means
of collecting and analyzing information that would be
equally accessible to any other user. In other words, it

 20
is a tool that allows journalists to see and understand
what is happening online – at a scale and with an
acuity that is not possible by looking at isolated
anecdotes.
But the scraping may not perfectly adhere to terms
and conditions required by each website, either
because perfect adherence would be impossible with
such a large number of websites or because the terms
of service of a website could change from one day to
the next to purport to prohibit data scraping without
a corresponding change to the relevant technological
protocols. Under the Eleventh Circuit’s broad reading
of the CFAA, however, this inadvertent overreach
might qualify as “misuse” of a computer system that
triggers criminal liability. An endorsement of this
interpretation by this Court would thus have an
enormous chilling effect on the automated
newsgathering techniques that have created so many
exciting opportunities for data journalists seeking to
create and analyze large datasets of publicly available
information.
In sum, data journalists like The Markup rely on
their ability to gather information from websites that
would be available to any other legitimate user, often
by means of automated technology. These basic and
fundamentally important reporting techniques would
be criminalized under the government’s and Eleventh
Circuit’s interpretation of the CFAA.

 21
III.

The First Amendment Protects the Online
Newsgathering
Practices
of
Data
Journalists.

Petitioner has explained (Pet. Br. 17-36) why the
court of appeals’ interpretation of the CFAA is wrong
as a matter of statutory interpretation. If there were
any question, though, the First Amendment would
prohibit that reading of the CFAA because it violates
the rights of journalists to gather and report on
publicly available information. The CFAA cannot,
consistent with the First Amendment, make it a crime
for The Markup or any other news organizations like
The Markup to gather publicly available information
from a website that they have lawfully accessed. Such
a result is prohibited by venerable decisions from this
Court protecting the right to gather and report upon
lawfully obtained information.
A. Routine Newsgathering Activities Cannot
Be
Criminalized
Under
the
First
Amendment Just Because They Were
Conducted Online.
News reporting is crucial to the proper functioning
of our self-governing democracy. Accordingly, the
First Amendment provides journalists with broad
protections against governments and individuals who
seek to thwart legitimate newsgathering activities.
While there are certainly limits on how far a reporter
can go to report a news story, this Court has time and
again made clear that journalists cannot be penalized
for gathering and publishing publicly available,
truthful information on a matter of public interest –
even in the face of laws requiring the information to
remain secret. An interpretation of the CFAA that

 22
enables website owners to make it a crime for
journalists to observe and report on information that
any other user could see would clearly violate these
deeply entrenched First Amendment principles. A
precedent effectively foreclosing newsgathering
online would also violate decisions of this Court
urging “extreme caution before suggesting that the
First Amendment provides scant protection for access
to vast networks” available through “the modern
Internet.” Packingham, 137 S. Ct. at 1736.
The decisions of this Court establishing the rights
of journalists to gather publicly accessible information
on a matter of public interest were developed in the
context of shoe-leather journalism, but are no less
applicable here. In Cox Broadcasting Corp. v. Cohn, a
reporter covering a rape-murder trial “learned the
name of the victim from an examination of the
indictments which were made available for his
inspection in the courtroom.” 420 U.S. 469, 472-73
(1975). He then broadcast the woman’s name in a
television news report, which indisputably violated a
Georgia statute that prohibited journalists from
publicizing the name of a rape victim. Id. But
application of the state privacy law violated the First
Amendment, this Court held, because “[o]nce true
information is disclosed in public court documents
open to public inspection, the press cannot be
sanctioned for publishing it.” Id. at 496. See also
Oklahoma Publ’g Co., 430 U.S. at 310 (holding that a
state court cannot “prohibit the publication of widely
disseminated information obtained at court
proceedings which were in fact open to the public”);
Garrison v. Louisiana, 379 U.S. 64, 74-75 (1964)
(“Truth may not be the subject of civil or criminal

 23
sanctions where discussion of public affairs is
concerned,” because such speech “is the essence of
self-government.”).
In a subsequent case, two newspapers learned the
identity of a juvenile charged with a shooting “by
monitoring the police band radio frequency” and
dispatching reporters to the scene, who “obtained the
name of the alleged assailant simply by asking
various witnesses, the police and an assistant
prosecuting attorney.” Smith v. Daily Mail Publ’g Co.,
443 U.S. 97, 99 (1979). Both newspapers published
the name of the alleged gunman and were indicted
under a West Virginia statute that made it a crime to
publish the name of a minor accused in a juvenile
proceeding. Id. at 100. Noting that the newspapers
were well within their rights to rely “upon routine
newspaper reporting techniques to ascertain the
identity of the alleged assailant,” this Court held that
“if a newspaper lawfully obtains truthful information
about a matter of public significance then state
officials may not constitutionally punish publication
of the information, absent a need to further a state
interest of the highest order.” Id. at 103-04 (holding
that the state’s interest in the confidentiality of
juvenile proceedings was not sufficient to overcome
First Amendment protections).
Perhaps most instructive is Florida Star v. B.J.F.,
491 U.S. 524 (1989). In that case, the Duval County
Sheriff’s Department inadvertently placed an incident
report containing the full name of a rape victim in the
precinct pressroom, where it was accessed by a
newspaper reporter. Id. at 526. The police did “not
restrict access to either the pressroom or to the reports

 24
made available therein” (id. at 527), but the
pressroom did display “signs making it clear that the
names of rape victims were not matters of public
record, and were not to be published.” Id. at 546
(White, J., dissenting). The newspaper ultimately
reported the rape victim’s name and was sued
successfully by the victim under a Florida statute that
prohibited disclosure. Id. at 529. In reversing the
jury’s verdict, this Court observed “that it is a limited
set of cases indeed where, despite the accessibility to
the public to certain information, a meaningful public
interest is served by restricting its further release by
other entities, like the press.” Id. at 535.
The Court also noted that the police department’s
failure to prevent access to the victim’s name did not
“make the newspaper’s ensuing receipt of this
information unlawful.” Id. at 536. To the contrary, the
fact that the reporter relied on “routine newspaper
reporting techniqu[es]” after the government made
the information publicly available “without
qualification, can only convey to recipients that the
government considered dissemination lawful.” Id. at
538-39. In other words, once the reporter was
authorized to access the rape victim’s name in the
press room, the government could not punish the
newspaper for publishing it – even though the police
should have kept it secret and even though there were
signs in the pressroom clearly telling the reporter that
the information was confidential. See also Bartnicki v.
Vopper, 532 U.S. 514, 535 (2001) (holding that
reporter who lawfully received and published contents
of illegally wiretapped phone call was not liable and
reaffirming earlier holdings that “state action to
punish the publication of truthful information seldom

 25
can satisfy constitutional standards”) (quoting Daily
Mail, 443 U.S. at 102).
These decisions all focus on the rights of news
organizations to publish information, but the
necessary corollary to the right to publish news is the
right to gather information in the first place. And each
of these cases tacitly affirms the core principle that
the government cannot punish journalists for using
“routine newspaper reporting techniques” to obtain
truthful information involving a matter of public
interest that has already been made publicly
available, even if it was disclosed by accident or
unlawfully. Daily Mail, 443 U.S. at 103-04.
This First Amendment protection applies even
when a journalist technically exceeds the terms under
which he or she was allowed to access the information.
See Florida Star, 491 U.S. at 534-35. For precisely the
same reasons, the First Amendment should protect
data journalists – including reporters working for The
Markup – who employ equally innocuous
newsgathering methods to gather truthful publicly
available data involving matters of public interest
from websites, even when the acquisition of that
information exceeds the terms under which they were
allowed to access it.6

This is not to say that website operators have no remedies
against users of their websites that violate the terms of service.
Indeed, causes of action for breach of contract and copyright
infringement may remain available in appropriate cases.
HiQLabs, Inc. v. LinkedIn Corp., 938 F.3d 985, 1004 (9th Cir.
2019). But what is not permitted is an “interpretation of the
CFAA’s ‘without authorization’ provision” that would “turn a
criminal hacking statute into a “sweeping Internet-policing
6

 26
An expansive reading of the CFAA to criminalize
journalists who violate a website’s terms and
conditions while in the process of gathering publicly
accessible information would also run afoul of this
Court’s longstanding commitment to preserve “the
vast democratic fora of the Internet.” Reno, 521 U.S.
at 868. “[C]yberspace … in general … and social media
in particular” are “essential venues for public
gatherings to celebrate some views, to protest others,
or simply to learn and inquire.” Packingham, 137 S.
Ct. at 1735 (emphasis added). Given the fact that this
Court has unequivocally extended the protections of
the First Amendment into the digital realm, it simply
cannot be the case that the First Amendment
protections afforded to journalists gathering public
information evaporate the moment they go online, if a
website operator should so choose.
More fundamentally, the Eleventh Circuit’s broad
reading of the CFAA undermines this Court’s efforts
to preserve the openness of the Internet as a source of
information. Indeed, this Court has urged “extreme
caution” before adopting any rules that would inhibit
“access to the vast networks” of the Internet because
“we cannot appreciate yet its full dimensions and vast
potential to alter how we think, express ourselves and
define who we want to be. The forces and directions of
the Internet are so new, so protean, and so far
reaching that courts must be conscious that what they
say might be obsolete tomorrow.” Packingham, 137 S.
Ct. at 1736. But the Eleventh Circuit’s broad
interpretation of the CFAA would effectively deprive
mandate” that would give website operators broad powers to
criminalize routine journalistic activities. Id. at 1000-01.

 27
journalists of the ability to engage in routine
newsgathering online. Such a reading would also
concentrate more power in the government and online
platforms by giving them an unprecedented veto
power – criminal sanctions – to punish anyone
attempting to investigate their activities. As society
increasingly migrates online, it is crucial to keep the
Internet open not just to the companies that profit
from it but also to the news organizations that seek to
tell the public how it works. Banning The Markup and
other journalists from collecting and analyzing
publicly accessible information would make it
impossible to observe activity on the Internet, which
in turn would make it impossible for reporters to serve
their vital oversight function.
In short, a broad reading of the CFAA that
criminalizes basic reporting practices used by data
journalists to obtain publicly accessible information
would clearly violate the First Amendment and must
be rejected.
B. The CFAA Should Be Interpreted to Avoid
Violating the First Amendment
Ultimately, this Court can avoid the confrontation
between the CFAA and the First Amendment by
interpreting it, as petitioner urges, not to criminalize
any access by authorized users even if they exceed
terms of service. “[I]t is a well-established principle
governing the prudent exercise of this Court’s
jurisdiction that normally the Court will not decide a
constitutional question if there is some other ground
upon which to dispose of the case.” Bond v. United
States, 572 U.S. 844, 855 (2014). Here, the

 28
constitutional question can be avoided simply by
giving the CFAA its most rational reading.
That was the course taken by the District Court for
the District of Columbia in Sandvig v. Barr, 2020 WL
1494065, at *1. In that case, academic researchers
tested job websites by creating “profiles for fictitious
job seekers” in order to determine whether job
rankings are “influenced by race, gender, age, or other
attributes.” Id. The researchers knew that their fake
profiles would violate the terms and conditions of the
hiring websites and brought a pre-enforcement
challenge, arguing that the CFAA violated their First
Amendment rights. Judge Bates recognized that an
interpretation of the CFAA that criminalized
academic research would raise First Amendment
concerns, but sidestepped this constitutional conflict
by interpreting the CFAA to “not actually criminalize
plaintiff’s proposed conduct – namely violating
website’s terms of service.” Id. at *7. The court
proceeded to interpret the statue and rejected the
Eleventh Circuit’s holding that a person violates the
CFAA by obtaining “information for a non-business
reason” in excess of their authorized access. Id. at *12
(quoting United States v. Rodriguez, 628 F.3d 1258,
1263 (11th Cir. 2010). Judge Bates noted that
“consumer websites’ terms of service do not provide
adequate notice for purposes of criminal liability”
because these “protean contractual agreements are
often long, dense, and subject to change.” Id. at *10,
*13. The court ultimately held that “violating public
websites’ terms of service … does not constitute a
CFAA violation under the ‘exceeds authorized access’
provision.” Id. at *13.

 29
*****
In sum, the CFAA cannot be interpreted to
criminalize routine newsgathering activities without
violating the First Amendment. But this
constitutional clash can be avoided simply by giving
the CFAA its rational interpretation, which takes into
account the ways in which information is gathered
online. The expansive interpretation of the CFAA that
the government advances here would criminalize
entirely lawful methods of gathering information on
the Internet and have a devastating chilling effect on
data journalism. Adopting petitioner’s interpretation
of the statute – which is consistent with the plain
meaning of the text, reflects the reality of how
newsgathering works online, and does not violate the
constitution – avoids this improper result.
CONCLUSION
For the foregoing reasons, the decision of the Court
of Appeals should be reversed.

Respectfully submitted,
Katherine M. Bolger*
John M. Browning
DAVIS WRIGHT TREMAINE LLP
1251 Avenue of the Americas
New York, NY 10020
(212) 489-8230
katebolger@dwt.com
jackbrowning@dwt.com

 30
David M. Gossett
DAVIS WRIGHT TREMAINE LLP
1919 Pennsylvania Ave. NW
Washington, DC 20006
(202) 973-4200
davidgossett@dwt.com
Nabiha Syed
THE MARKUP
900 Broadway, Suite 202
New York, NY 10159
(347)894-3746
nabiha@themarkup.org
* Counsel of Record
Counsel for Amicus Curiae
July 8, 2020