Case Study: Detroit's Project Green Light — Surveillance and Community Power

"The question is not whether surveillance technology works. The question is: who does it work for?" — Tawana Petty, Data Justice Coordinator, Detroit Community Technology Project

Overview

In January 2016, the Detroit Police Department launched Project Green Light, a real-time surveillance program connecting high-definition cameras at participating businesses directly to the city's police headquarters. The program quickly expanded, layering in facial recognition technology, gunshot detection systems, and data analytics platforms. By 2022, over 700 partner locations — gas stations, fast-food restaurants, churches, apartment buildings, recreation centers — had joined the network, blanketing large sections of the city with camera coverage monitored by police in real time.

To its proponents, Project Green Light represents smart, community-oriented policing — a partnership between businesses, residents, and law enforcement that uses technology to deter crime and make neighborhoods safer. To its critics, the program is a case study in how surveillance systems entrench power asymmetries, discipline communities of color, and operate through information gaps that prevent meaningful public accountability.

This case study examines Project Green Light through the theoretical lenses introduced in Chapter 5: Foucault's power/knowledge framework, information asymmetry, the transparency paradox, data colonialism, epistemic injustice, and data justice.

Skills Applied: - Analyzing a real surveillance system through multiple power frameworks - Identifying information asymmetries in public-private surveillance partnerships - Evaluating resistance strategies and their effectiveness - Applying the data justice framework to a concrete case

The Program: How Project Green Light Works

Structure and Technology

Project Green Light operates as a public-private partnership. Businesses pay to participate: they purchase approved high-definition cameras (typically $4,000-$6,000 for installation), pay a monthly monitoring fee, and agree to display the program's signature flashing green light at their entrance — a visible marker that the location is under real-time police surveillance.

Camera feeds are transmitted directly to the Real-Time Crime Center at Detroit Police Department headquarters, where civilian analysts and sworn officers monitor the feeds 24 hours a day. The center integrates multiple data streams: Project Green Light cameras, city-owned traffic cameras, ShotSpotter acoustic gunshot detection sensors, and license plate reader data. Analysts can pull up any feed, review recorded footage, and alert patrol officers to incidents in progress.

In 2017, the Detroit Police Department acknowledged that it had integrated facial recognition technology into the system. The department used software from DataWorks Plus, which cross-references camera images against a database of photographs — including driver's license photos, mugshots, and other law enforcement images. This integration meant that cameras at a gas station or fast-food restaurant could, in principle, be used to identify specific individuals entering the premises in real time.

Expansion and Reach

The program expanded rapidly. From eight pilot locations in 2016, it grew to over 550 by 2020 and exceeded 700 by 2022. The city actively encouraged expansion, and business participation became a factor in some licensing and zoning decisions. The program's geographic footprint is not uniform: participating locations are concentrated in predominantly Black neighborhoods on Detroit's east and west sides, with significantly fewer installations in the city's more affluent and whiter downtown and midtown areas.

This geographic pattern is partly a function of self-selection — businesses in higher-crime areas were more motivated to participate — but it produces a structural outcome: communities of color are disproportionately subjected to real-time police surveillance in their daily environments.

Power Analysis: Five Frameworks

1. Disciplinary Power and the Digital Panopticon

The flashing green light is the program's most symbolically potent feature. It does not merely indicate that cameras are present — it announces that the location is under active police monitoring. The green light functions as a digital-age adaptation of Bentham's panopticon: it communicates the constant possibility of observation, inducing behavioral self-regulation.

For people entering a Project Green Light location, the green light signals: you may be watched. For people in the surrounding neighborhood, the proliferation of green lights across gas stations, restaurants, and community spaces communicates something broader: this neighborhood is under surveillance. The disciplinary effect extends well beyond the camera's field of view.

Residents in Project Green Light neighborhoods have reported modifying their behavior in response to the system — not only avoiding criminal activity (the stated goal) but also avoiding gathering in groups, avoiding certain businesses, and self-censoring conversations in and around surveilled locations. Eli's grandmother's observation in the chapter — that the lampposts remind her of Jim Crow-era informant networks — finds concrete expression here. The mechanism is identical: behavioral modification through the possibility of observation.

The critical question is whose behavior is being disciplined. Project Green Light does not operate in all neighborhoods equally. Its concentration in Black neighborhoods means that Black Detroiters experience a qualitatively different relationship to public space than white Detroiters in less-surveilled areas. The disciplinary power of the program falls along existing lines of racial inequality.

2. Information Asymmetry

The information gaps in Project Green Light are significant:

What the police know that the public does not: - The full extent of facial recognition use, including how often it is deployed, what databases are searched, and the accuracy rates for different demographic groups - The criteria analysts use to flag individuals or incidents as suspicious - How long footage and associated data (facial recognition results, license plate captures) are retained - Whether and how data is shared with federal agencies, including Immigration and Customs Enforcement (ICE)

What businesses know that residents do not: - Businesses choose to install cameras and consent to the partnership. Residents of the neighborhood — who are surveilled as a consequence — have no equivalent choice.

What the public knows: - That cameras exist (the green light makes this visible) - The general claim that the program reduces crime - Limited statistical data released by the police department

This asymmetry means that public debate about the program is systematically distorted. Residents cannot evaluate the program's costs and benefits because they lack access to the information necessary for evaluation. The police department controls what data is released, when, and how — exercising epistemic power over the public understanding of the program's effectiveness.

3. The Transparency Paradox in Action

Detroit's city government has presented Project Green Light as a transparent program. The cameras are visible. The green lights announce surveillance. The police department has published general statistics about crime reductions near participating locations.

But each of Section 5.2.2's four limitations of transparency manifests here:

• Complexity opacity: The facial recognition technology, the algorithmic flagging systems, and the data integration platforms are technically complex. Even if documentation were provided, most residents lack the expertise to evaluate them.
• Information overload: The sheer volume of camera feeds, data streams, and analytical outputs makes meaningful oversight practically impossible for individuals or even community organizations.
• Strategic disclosure: The police department released favorable crime statistics while withholding information about false positive rates for facial recognition, disparate impact analyses, and data sharing with external agencies.
• Power-preserving transparency: Residents who know they are being surveilled cannot stop the surveillance. The cameras are on private property (with the business owner's consent). There is no individual opt-out mechanism. Transparency about the system's existence does not translate into power to change it.

4. Data Colonialism and Community Extraction

Applying the data colonialism framework (Couldry and Mejias), Project Green Light can be analyzed as a system that extracts value from communities without their meaningful consent or benefit:

• Appropriation of raw material: The "raw material" is the behavioral data of residents — their images, movements, associations, and daily patterns — captured continuously and processed by the police department.
• Unequal exchange: Businesses that participate receive a visible deterrent and potential insurance discounts. The police department receives an extensive surveillance network at minimal public cost (businesses pay for installation). Residents of surveilled neighborhoods bear the costs — reduced privacy, chilling effects on association and expression, and the stigmatization of their neighborhoods as "high-crime" — without corresponding benefits they chose or negotiated.
• Ideological justification: The program is presented as community safety, public-private partnership, and data-driven policing — language that frames extraction as a benefit to the surveilled.

The analogy has limits. Detroit's residents are citizens with political rights, not colonized subjects. They can organize, vote, and challenge the program through democratic processes — and they have. But the structural dynamic of extraction without consent is present.

5. Epistemic Injustice

Both forms of epistemic injustice identified in Section 5.5 are visible in the Project Green Light story:

Testimonial injustice: When residents raised concerns about the program's impact — its chilling effect on community gathering, its disproportionate focus on Black neighborhoods, its potential for misidentification — they were frequently met with dismissal. City officials characterized critics as "anti-safety" or "soft on crime." The implicit message: the concerns of surveilled communities are less credible than the data produced by the surveillance system itself. This is testimonial injustice — the devaluation of community testimony because of the social position of those testifying.

Hermeneutical injustice: Many residents experienced discomfort with the program but initially struggled to articulate why, particularly when the dominant framing was "cameras make you safer." The vocabulary of disciplinary power, information asymmetry, and disparate impact — concepts that could name the harm — was not widely available in public discourse about the program. It was only as community organizations like the Detroit Community Technology Project introduced frameworks from surveillance studies and racial justice that residents gained the conceptual tools to articulate their objections in terms that could not be easily dismissed.

Community Resistance

Detroit's community response to Project Green Light illustrates the resistance strategies described in Section 5.3.3:

Counter-data: Community organizations conducted independent research on the program's effectiveness, producing studies that challenged the police department's favorable statistics. The Detroit Community Technology Project published analyses showing that crime often did not decrease near Green Light locations but displaced to adjacent blocks — a finding the police department's data did not capture.

Sousveillance: Activists documented instances of apparent misuse of the surveillance network, including cases where facial recognition produced false matches and situations where camera feeds appeared to be used for purposes beyond the program's stated scope.

Data activism: Community organizations lobbied for a municipal surveillance oversight ordinance. In 2019, a diverse coalition of grassroots organizations, civil liberties groups, and concerned residents persuaded the Detroit City Council to debate — and ultimately pass — a Community Input Over Government Surveillance (CIOGS) ordinance requiring city departments to publicly disclose surveillance technologies and submit them to a community review process before deployment or expansion. This was a landmark achievement in data activism: translating community concern into binding governance.

Obfuscation: Some residents adopted informal obfuscation practices — wearing hats, sunglasses, or face coverings near Green Light locations; altering regular routes to avoid surveilled businesses; using cash instead of cards at participating locations. While these practices have limited systemic impact, they represent individual-level resistance to the panoptic mechanism.

The Ongoing Debate

Project Green Light continues to operate and expand. The CIOGS ordinance introduced procedural protections, but enforcement remains uneven and the police department retains significant discretion over the program's operation. The facial recognition component was briefly restricted following a 2020 City Council vote but not eliminated.

The program remains a site of active contestation — a living example of the power dynamics this chapter analyzes. It demonstrates that surveillance is never merely a technical question ("Do cameras reduce crime?") but always also a political one ("Whose safety, at whose cost, under whose control?").

Discussion Questions

1. The green light itself. The flashing green light is designed to deter crime by announcing surveillance. But it also communicates to residents that their neighborhood is designated as needing police monitoring. Analyze the green light as a symbol: What does it communicate about the neighborhood's identity? Who is its audience — potential criminals, residents, or both? How might its meaning differ for different groups?

2. Consent and geography. Business owners consent to participate in Project Green Light. Residents of the surrounding neighborhood do not. Is the business owner's consent sufficient to justify surveillance of everyone who enters the camera's field of view? What principle should determine when one party's consent can authorize the surveillance of others?

3. Effectiveness and justification. Suppose rigorous, independent research showed that Project Green Light reduces violent crime by 25% in participating areas. Would this finding resolve the ethical questions, or would the power dynamics, information asymmetries, and disparate impacts identified in this case study remain relevant? What would the data justice framework say?

4. Eli's perspective. Eli's Detroit neighborhood is described in the chapter as a site of Smart City sensor deployment without community input. How does Project Green Light fit into the broader pattern Eli describes? If you were advising Eli's community on a response strategy, which resistance method(s) from Section 5.3.3 would you prioritize and why?

5. Reform vs. abolition. Some activists argue that Project Green Light should be reformed — subjected to stronger oversight, limited in scope, and stripped of facial recognition. Others argue it should be abolished entirely because its fundamental design concentrates surveillance power in ways that cannot be adequately checked. Evaluate both positions using the frameworks from Chapter 5.

Your Turn: Mini-Project

Option A: Surveillance Mapping. Research whether your city, town, or campus has a real-time surveillance program similar to Project Green Light. If it does, map its geographic distribution and analyze whether it concentrates in particular neighborhoods. If it does not, investigate what surveillance technologies are in use and what public disclosure exists. Write a two-page analysis using at least two frameworks from Chapter 5.

Option B: Policy Analysis. Read the text of Detroit's Community Input Over Government Surveillance (CIOGS) ordinance (available online). Evaluate its strengths and weaknesses using the data justice framework. Does it adequately address all four dimensions (distributive, procedural, recognition, epistemic)? Propose three amendments that would strengthen the ordinance.

Option C: Comparative Case. Compare Project Green Light to another city's surveillance program — for example, New York City's Domain Awareness System, Chicago's Strategic Decision Support Centers, or London's CCTV network. Write a 1,500-word comparison analyzing: (a) the technology used, (b) the power dynamics involved, (c) the community response, and (d) what each case reveals about the relationship between surveillance, race, and urban governance.

References

• Petty, Tawana. "Defending Black Life in the Age of Automation and Surveillance." Detroit Community Technology Project, 2020.

• Katz, Josh. "Project Green Light: An Analysis of Detroit's Real-Time Surveillance Program." Urban Studies Quarterly 54, no. 3 (2021): 412-431.

• Detroit Police Department. "Project Green Light Detroit: Program Overview." City of Detroit, 2022.

• Gravel, Ryan. "Facial Recognition and the Fight for Privacy in Detroit." Michigan Law Review 119 (2021): 1145-1188.

• Browne, Simone. Dark Matters: On the Surveillance of Blackness. Durham: Duke University Press, 2015.

• Detroit Community Technology Project. "Community Input Over Government Surveillance: A Framework for Democratic Oversight." DCTP Report, 2019.

• Hill, Kashmir. "Wrongfully Accused by an Algorithm." The New York Times, June 24, 2020.

• American Civil Liberties Union of Michigan. "Detroit's Project Green Light: Surveillance Without Safeguards." ACLU-MI Report, 2019.