A Critical Summary of Detroit’s Project Green Light
and its Greater Context
Detroit Community Technology Project
June 9, 2019

Research provided by Noah Urban, Jacob Yesh-Brochstein, Erica Raleigh and Tawana Petty

 Contents
What is Project Green Light? ......................................................................................... 3 
History ........................................................................................................................... 3 
System Structure .......................................................................................................... 3 
Facial Recognition ...................................................................................................... 4 
Expansion ..................................................................................................................... 5 
What does the Community Think? ............................................................................... 5 
Why is Project Green Light controversial? ................................................................... 6 
Controversial Effectiveness ........................................................................................ 6 
Controversial Policies .................................................................................................. 7 
Controversial Results ................................................................................................... 8 
Controversial Costs ..................................................................................................... 9 
Controversial Path ...................................................................................................... 9 
Two Paths Forward ....................................................................................................... 10 
References .................................................................................................................... 11 

2

 What is Project Green Light?
History
According to Captain Sloan of the Detroit Police Department (DPD), Project
Green Light (PGL) was conceived after DPD observed that 25% of violent crime
in the city was occurring within 500 feet of gas stations. DPD subsequently
performed risk modeling that reportedly concluded that late night businesses
such as gas stations and party stores were at the greatest risk for violent crime (1;
2).
PGL was first implemented on January 1st, 2016, as a partnership between local
business, the City of Detroit, and community groups. On that day, eight gas
stations had real-time security cameras connected with DPD headquarters for
monitoring (3). In the three and a half years since then, PGL has expanded
rapidly. As of June 2019, there are over 550 participating locations in the city,
ranging from gas stations, to churches, to reproductive health centers (4).

Map of Detroit showing Project Green Light locations and their increase over 4 years. Credit: Cyrus Peñarroyo

System Structure
Camera positioning is determined by DPD in coordination with the City of
Detroit. Outdoor cameras cover all areas generally accessible to the public on
and around the participants property. They are positioned to be able to capture
3

 license plates and avoid glare from light sources. For indoor cameras, there must
be at least one camera directly facing regularly used entrances. Each
participant of PGL must spend money to become involved in the program. They
must purchase and install cameras, PGL signage, decals, and green lights to
signify to customers that they are part of the project (3).
PGL cameras currently send live feeds to a 24-hour crime monitoring surveillance
center called a “real time crime center” (RTCC). The RTCC opened in
November 2017 and is operated by both sworn officers and unsworn individuals.
The RTCC has an array of surveillance tools at their disposal including facial
recognition software, license plate readers, and access to GPS tether data. They
receive other feeds in addition to PGL, including feeds from Rock Financial’s
cameras, hidden neighborhood cameras (“originally put in place to catch
dumping”). The RTCC works with government partners including the FBI and the
Department of Homeland Security, and private partners including, DTE, Rock
Financial, Downtown Detroit Partnership, and the Ilitchs (1; 2).
The real-time crime center monitors incidents at PGL location in real time
through the cameras. They “virtually patrol” each PGL location once a day and
physically patrol around 184 PGL locations each day. Around 30 calls from PGL
locations come through the center each day. Additionally, PGL locations are
given priority over non-PGL locations (1; 2).
Facial Recognition
The City of Detroit put forth a Request for Proposals for a contractor to work
closely with the city, DPD, and Motorola (Company that help set up the RTCC)
to set up a “turn-key” facial recognition system that would work with the already
existing infrastructure of the RTCC. They specifically asked that the facial
recognition work on at least 100 concurrent real-time video feeds, be integrated
into the PGL system, and can be used by officers with a mobile app (5).
The city received 3 proposals, accepting the one from DataWorks Plus. In late
2017, the city signed a 3 year contract with them using over $1,040,000 in city
funds. The city purchased the Face Plus facial recognition solution, toted by
DataWorks Plus to use a combination of two algorithms (ROC and NEC) to
match faces citing its program being used by JNET (Pennsylvania), Michigan
State Police, San Bernardino County, Chicago Police, and Riverside County.
Face Plus is capable of automatically searching all faces that enter camera
frames against photos in the entity’s database, alerting authorities to any
algorithmic matches. Additionally, there is a “watchlist” option where persons of
interest can be monitored and alerted for (5).

4

 DPD currently uses SNAP photos for their database. SNAP photos include mug
shots, sex offender registry photographs, driver’s license photos and state ID
photos (6). According to DataWorks Plus, in 2017, this repository contained 8
million criminal pictures and 32 million “DMV” pictures (5). As the Free Press
reported in March 2019, almost every Michigan resident has a photo of them in
this system (7).
Expansion
Claiming PGL promotes both economic development and safety, the City of
Detroit plans to continue extending the program to entities other than
commercial businesses (restaurants, multi-family dwellings, mixed use
developments, etc.) (1; 2; 8). Despite research contradicting camera presence
in schools effectiveness and possible psychologically detrimental effects, an
expansion into public schools has already begun – 23 cameras in Randolph
Career Technical Educational Center are being monitored at the RTCC (9).
According to the DPD, the superintendent will be working on getting all schools
into PGL (1; 2; 10). This has already led to students protesting PGL (11).
In the Mayor’s proposed Capital Agenda 2020-2024, Mike Duggan called for a
$1,100,000 PGL expansion in 2019, along with a $350,000 UAV (Drone) program
expansion. Additionally, he called for $4,000,000 by 2020 to expand the RTCC
and create two “mini” RTCCs. Duggan also proposed adding PGL to all public
transit stops, in buses, and to integrate facial recognition in the streets through a
pilot of MioVision Traffic Signal Cameras (12).

What does the Community Think?
For the past four years, the Detroit Community Technology Project (DCTP) has
co-led a participatory research project called the Our Data Bodies Project
(ODB). Residents across three cities – Detroit, Los Angeles and Charlotte – were
asked questions regarding the impact of how their data is used and how they
experience surveillance technologies.
Through research conducted with over 130 residents between the three cities,
ODB discovered that surveillance and data collection was deeply connected to
diversion of public benefits, insecure housing, loss of employment opportunities,
and the policing and subsequent criminalization of the community members
that come into contact with these surveillance systems.

5

 The research was conducted to answer three main questions:
1. How do marginalized adults experience and make sense of the
collection, storage, sharing and analysis of their personal information?
2. How, if at all, do marginalized adults connect their ability to meet their
basic material and social needs to their inclusion in (or exclusion from)
data-based systems?
3. What strategies do marginalized adults deploy, if any, to protect their
digital privacy, self-determination, and data rights?
Across all three cities, community members expressed an interest in being seen,
but indicated that they did not want to feel like they were being watched. They
wanted both privacy and the ability to be seen and heard as whole human
beings. One resident interviewed said, “I mean your face is not even your own
anymore. Your face is being captured by cameras.”
ODB’s interim report was recently read into congressional record by Brandi
Collins-Dexter, Senior Campaign Director for Color of Change (13).
Congresswoman Rashida Tlaib (MI-13), representing a large portion of Detroit,
spoke on facial recognition technology during a House Oversight Committee
hearing recently. Expressing her anxiety around the technology she stated, “This
stuff freaks me out." (14)
The Detroit Digital Justice Coalition (DDJC) through its coalition member Detroit
Community Technology Project (DCTP), has joined the growing number of fellow
Detroiters concerned or opposed to the controversial expansion of Project
Green Light and related facial recognition technologies (15).

Why is Project Green Light controversial?
The idea that public surveillance equates to public safety has led local
governments to make problematic decisions that facilitate an outcome that is
either ineffective or unsafe (16). Such policies, such as predictive policing, may
disproportionately affect marginalized peoples (undocumented, formerly
incarcerated, unhoused, poor, etc.) and minority (black, latinx, etc) populations
(17).
Controversial Effectiveness
Areas in Detroit and other cities that are inundated with poverty due to
decades of disinvestment are subject to higher crime rates (18). While mass
6

 surveillance and more militarized policing may be the route that certain
municipalities may be going in – despite evidence contradicting its efficacy – it
has been argued that reinvestment into the community that does not displace
or further marginalize residents (through housing, education, affordable water,
etc.) would actually be effective in reducing crime and improving public safety.
This is supported by research that has shown that crime rates increase when
public welfare is unable to support individuals (19; 20).
Although many public agencies frequently discuss the benefits of programs such
as PGL, past studies have shown that the actual results of these programs vary
widely. Other closed-circuit surveillance projects, which are not a direct analog
for PGL, have seen mixed impact. In neighborhoods in Chicago, Baltimore, and
Newark for example, some areas saw substantial reductions in crime while others
did not (21).
Another thing to think about regarding PGL is that even if a positive effect could
be proven, it would be diminished upon an increase in scale. The Memorandum
of Understanding (MOU) PGL partners sign with the City states that DPD is not
obliged to monitor the cameras at all times. (8) Rapid expansion of the program
may lead to even less camera monitoring as resources become too strained.
This is especially the case if the program grows too quickly – and the
department has acknowledged this (1; 2). Researchers at the Urban Institute
agree that such a scenario could become quickly problematic, stating that it’s
quite easy to, “get to the point where it’s so saturated that it becomes
ineffective for everybody.” That is, even if PGL could be proven to be an
effective deterrent for crime, it will likely lose its effectiveness if it is believed that
the cameras are not being monitored (21).
Additionally, costs will increase to sustain a larger program (21). As costs
increase, it may be worth questioning if the costs of reactionary policies like
massive surveillance are as effective in reducing crime as spending similarly
large quantities on preventative policies such as increases in public goods.
Controversial Policies
DPD recently released two policy drafts regarding the use of facial recognition
and traffic light mounted camera footage (22; 6). There were some points it
contained (and didn’t contain) that are of concern to some residents:
-

While policies state that DPD cannot use the cameras for immigration
enforcement purposes, the policies do not explicitly exclude DPD from
letting the Department of Homeland security (a close partner) from
accessing footage or footage related information.
7

 -

Non-law enforcement entities, (like Rock Financial) may submit an image
for a facial recognition search after filing an incident report.
Facial recognition may be used (upon approval of DPD officials) to
monitor “first amendment-protected events”.
Facial recognition information may be used not just for criminal
investigation, but “other types of inquiries” as well.
If someone is suspected of a crime, their face will be stored in the system
until they are identified, no longer a suspect, or the statute of limitation
has expired.
The public may or may not be notified of data breaches
One individual, the person in charge of the facial recognition program,
will also be the only one responsible for overseeing complaints of bias
against the program and responding to them.

Map of Detroit showing the relationship between Project Green Light locations and average police response times.
Credit: Cyrus Peñarroyo

Controversial Results
There is substantial political will to expand PGL, especially from the Mayor and
the chief of police. There are many statistics used as evidence for PGL’s
effectiveness.
•

Incidents of violent crime reduced by 48% (compared to 2015) at the
original 8 PGL sites (3).

8

 •
•

Incidents of violent crime reduced by 23% (compared to 2015) at all PGL
sites (3).
Car-jackings have decreased by 40% in 2 years (23).

These statistics have been deemed too short of a sample size and/or too short of
a time frame by researchers (11). It has also been noticed, and confirmed by
the Police Chief, that there have been no studies comparing PGL locations to
non-PGL locations. In the absence of this, researchers say it is nearly impossible
to tie Detroit's crime reduction specifically to Green Light. Additionally, the stats
are in line with crime rates that have been decreasing since at least 2007,
according to the FBI (24). Motor vehicle theft fell over 30% between both 20072009 and 2013-2015. From 2009-2012 violent crime fell by almost 17%.
Controversial Costs
In January 2018, the mayor announced plans to mandate that all retail business
with late night hours partake in PGL, initially costing each business $1,000 to
$6,000 with an additional yearly charge of over $1,600 a year in video storage
(23; 21). Already, some business owners in the city have confided to reporters
that they feel like they are paying for policing (25). Separately, officials such as
Commissioner Burch and Deputy Chief Bettison have threatened to
decommission PGL locations that do not meet certain beautification standards,
in effect removing their priority 911 status stating, “We will pull that light right off
of that business and move forward” (1; 2).
Controversial Path
Many cities have started out with surveillance programs and then moved on to
more aggressive programs such as predictive policing, drones, and facial
recognition. Detroit is currently on track to follow such a trajectory – it already is
utilizing facial recognition for PGL. It plans on using facial recognition on soon to
be installed traffic cameras. Such programs have been surrounded by
controversy including racial bias, so much so, that some like Los Angeles have
even backtracked the polices. There is legitimate fear regarding what that
would look like in a majority black city such as Detroit (26; 27; 28; 29).
Some cities, like San Francisco, have preemptively decided that this is a path
they do not want to take, banning the use of aggressive surveillance
technologies like facial recognition (30). Other cities like Baltimore, have
increased its frequency and scale of use, subsequently coming under scrutiny
for inappropriate use and racial bias. One example of such use occurred in
2015, when Baltimore police used facial recognition to arrest anyone with
outstanding warrants who were protesting the death of Freddie Gray (31). Such
9

 use would be sanctioned as well under Detroit’s facial recognition policy (6).
There is growing concern that the use of facial recognition software used at
protests will lead to watch lists and intimidation, thereby infringing on the human
right to protest (32).

Two Paths Forward
At this juncture in Detroit’s history, there are two clear paths forward. The first
path sees more surveillance, predictive policing, facial recognition, and other
privacy compromising practices with, at the very least, questionable impacts on
preventing crime (33). It sees more jails (34). It centers on the belief that people
are the problem and that the problem can be solved through fear of
punishment, punishment, and removal of said people from society. Such a path
could very well lead to greater marginalization of already marginalized
communities (21).
The second path is different. It does not assume that people are the main cause
of crime. The second path is grounded in the research that shows that the
conditions people live in are the greatest influence on criminal activity. It
reduces crime by addressing inequalities and societal shortcomings via
investment in people and their communities (35). It guards privacy while
increasing trust in the government. It both protects and serves.

10

 References
1. City of Detroit. Detroit Police Commissioners Meeting 04 19 2018. DetroitMI.gov.
[Online] April 19, 2018. [Cited: June 5, 2019.]
http://video.detroitmi.gov/CablecastPublicSite/show/5963?channel=3.
2. —. Detroit Board of Police Commissioners Regular Meeting. DetroitMI.gov.
[Online] April 19, 2018. [Cited: June 5, 2019.]
https://detroitmi.gov/Portals/0/docs/policecommissioners/BOPC%20Minutes%20
%20%2004-19-18.pdf?ver=2018-05-03-121332-883.
3. —. Project Green Light Detroit. [Online] 2019. [Cited: June 4, 2019.]
https://detroitmi.gov/departments/police-department/project-green-lightdetroit.
4. Detroit Police Department. Project Green Light Locations. City of Detroit Open
Data Portal. [Online] June 5, 2019. [Cited: June 5, 2019.]
https://data.detroitmi.gov/Public-Safety/Project-Green-Light-Locations/vrh66pvj.
5. City of Detroit. Professional Services Contract Between City of Detroit,
Michigan and DataWorks Plus Contract No. 6000801. [Online] 2017. [Cited: June
5, 2019.]
https://hub.novatuscontracts.com/novatus/services/detroit/downloadDocume
nt?documentId=64c55d4f-deb2-4423-94db-62fc2e1e892d.
6. —. Facial Recognition Directive 307.5. DetroitMI.gov. [Online] January 2019.
[Cited: June 2019, 5.] https://detroitmi.gov/document/facial-recognition.
7. Egan, Paul. Never arrested? Michigan State Police still likely has your photo in
its database. Detroit Free Press. [Online] March 11, 2019. [Cited: 5 2019, June.]
https://www.freep.com/story/news/local/michigan/2019/03/11/michigan-statepolice-facial-recognition-database/3102139002/.
8. City of Detroit. Memorandum of Understanding. [Online] 2019. [Cited: June 4,
2019.] https://detroitmi.gov/departments/police-department/project-greenlight-detroit/agreements.
9. Surveillance or Safekeeping? How School Security Officer and Camera
Presence Influence Students’ Perceptions of Safety, Equity, and Support.
Lindstrom Johnson, Sarah, et al. 6, s.l. : journal of Adolescent Health, December
2018, Vol. 63, pp. 732-738.

11

 10. City of Detroit. DETROIT BOARD OF POLICE COMMISSIONERS REGULAR
MEETING 4-5-18. DetroitMI.gov. [Online] April 5, 2018. [Cited: June 5, 2019.]
https://detroitmi.gov/Portals/0/docs/policecommissioners/BOPC%20Minutes%20
04-05-18.pdf?ver=2018-05-03-120540-760.
11. Ikonomova, Violet. Detroit students fight Project Green Light's expansion to
schools. Metro Times. April 23, 2018.
12. City of Detroit. City of Detroit Proposed Capital Agenda 2020-2024.
DetroitMI.gov. [Online] November 2018. [Cited: June 5, 2019.]
https://detroitmi.gov/sites/detroitmi.localhost/files/201811/City%20of%20Detroit%20Proposed%20Capital%20Agenda%2020202024%20Full_1.pdf.
13. Collins-Dexter, Brandi. Testimony of Brandi Collins - Dexter. House.gov.
[Online] February 26, 2019. [Cited: June 6, 2019.]
https://energycommerce.house.gov/sites/democrats.energycommerce.house.
gov/files/documents/Brandi%20Collins%20Dexter%2002.26.2019.pdf.
14. Johnson, Jake. ‘This stuff freaks me out’: Rep. Rashida Tlaib raises alarm over
use of facial recognition as groups demand federal moratorium. Raw Story.
[Online] June 4, 2019. [Cited: June 6, 2019.]
https://www.rawstory.com/2019/06/this-stuff-freaks-me-out-rep-rashida-tlaibraises-alarm-over-use-of-facial-recognition-as-groups-demand-federalmoratorium/.
15. University of Michigan: Semester in Detroit. Detroiters Speak 3/5/2019:
Containment & Surveillance - Shifting Borders & Boundaries. Youtube. [Online]
April 16, 2019. [Cited: June 6, 2019.]
https://www.youtube.com/watch?v=IEXqb3bgoAk.
16. Guariglia, Matthew. Too much surveillance makes us less free. It also makes
us less safe. The Washington Post. July 18, 2017.
17. Shapiro, Aaron. Reform predictive policing. Nature. January 25, 2017, Vol.
541, pp. 458-460.
18. Sackett, Chase. Neighborhoods and Violent Crime. Evidence Matters. 2016,
Vol. Summer.
19. Welfare Payments and Crime. Foley, C. Fritz. 1, February 2011, The Review of
Economics and Statistics, Vol. 93, pp. 97-112.
20. Lafree, Gary. Losing Legitimacy: Street Crime And The Decline Of Social
Institutions In America. 1st. New York : Routledge, 1998.
12

 21. Gross, Allie. Controversial surveillance program coming to Detroit public
housing. Detroit Free Press. November 6, 2018.
22. City of Detroit. Use of Traffic Light-Mounted Cameras and Facial Recognition
Technology Directive 307.6. DetroitMI.gov. [Online] April 23, 2019. [Cited: June 5,
2019.] https://detroitmi.gov/document/use-traficc-light-mounted-cameras-andfacial-recognition-technology.
23. Livengood, Chad. Detroit aims to mandate Project Green Light crimemonitoring surveillance for late-night businesses. Crain's Detroit Business. January
3, 2018.
24. FBI. Crime Data Explorer. [Online] 2019. [Cited: June 4, 2019.] https://crimedata-explorer.fr.cloud.gov/explorer/agency/MI8234900/crime/2007/2017.
25. Ikonomova, Violet. Project Green Light faces scrutiny as Detroit eyes
mandate for thousands of businesses. Detroit Metro Times. [Online] January 17,
2018. [Cited: June 6, 2019.] https://www.metrotimes.com/detroit/project-greenlight-faces-scrutiny-as-detroit-eyes-mandate-for-thousands-ofbusinesses/Content?oid=8539251.
26. Wynter, Kareen. LAPD Chief Moore Announces Ending Data-Driven Crime
Prevention Program. KTLA 5 News. April 9, 2019.
27. Neavling, Steve. Researchers alarmed by Detroit's pervasive, expanding
facial-recognition surveillance program. Detroit Metro Times. May 17, 2019.
28. Stop LAPD Spying Coalition. Dismantling Predictive Policing in Los Angeles.
Los Angeles : s.n., 2018.
29. Newman, Lily Hay. Facial Recognition Tech is Creepy When it Works - And
Creepier When it Doesn't. Wired. [Online] May 9, 2018. [Cited: June 6, 2019.]
https://www.wired.com/story/facial-recognition-tech-creepy-works-or-not/.
30. Conger, Kate, Fausset, Richard and Kovaleski, Serge F. San Francisco Bans
Facial Recognition Technology. The New York Times. May 14, 2019.
31. Baltimore Sun Editorial Board. Bias in facial recognition technology needs to
be fixed. The Baltimore Sun. [Online] June 3, 2019. [Cited: June 6, 2019.]
https://www.baltimoresun.com/opinion/editorial/bs-ed-bias-bias-facialrecognition-technology-20190524-story.html.
32. Hashem, Hoda. Police Use of Facial Recognition Surveillance Could Violate
our Rights. Rights Info. [Online] September 24, 2018. [Cited: June 6, 2019.]
https://rightsinfo.org/police-use-of-facial-recognition-a-threat-to-human-rightsand-probably-unlawful/.
13

 33. Evelth, Rose. In the Face of Danger, We're Turning to Surveillance. Wired.
[Online] March 21, 2019. [Cited: June 6, 2019.]
https://www.wired.com/story/surveillance-safety/.
34. Corey, Samuel. Detroit teens reimagine criminal justice with alternative to
new jail. Detroit Metro Times. October 5, 2018.
35. Wilkinson, Richard, [perf.]. How Economic Inequality Harms Societies. [prod.]
www.TED.com. Youtube.com, 2011.

 

14