The Neighborhood Watch Goes Corporate: How Amazon and Google Are Selling Mass Surveillance to American Cities — and Who Pays the Price
In 2023, the Los Angeles Police Department maintained active data-sharing partnerships with more than 2,000 Amazon Ring cameras operated by private residents — a network assembled not through any city council vote, legislative authorization, or public hearing, but through a corporate program that Amazon had quietly extended to law enforcement agencies across the country. At its peak, Ring's law enforcement portal — a platform called Neighbors — had formalized partnerships with more than 2,400 police departments in the United States, giving officers the ability to request footage from residential cameras without a warrant, sometimes without the knowledge of the camera's owner.
That program, exposed in detail by investigative journalists at The Intercept and later scrutinized by a Senate inquiry, offers a window into something that has been happening in American cities for years with almost no public reckoning: the systematic construction of a surveillance infrastructure built on corporate hardware, funded by municipal contracts, and deployed with a consistency that tracks not crime rates but demographics.
The Infrastructure of Watching
The modern urban surveillance apparatus is not a single system. It is a layered accumulation of technologies that, taken together, create something qualitatively different from any prior conception of public monitoring. Ring and Nest cameras cover residential streets and front doors. License plate readers — sold by companies like Motorola and Axon — log the movement of vehicles through intersections and highways, creating databases of location history for millions of drivers who have committed no offense. Facial recognition software, increasingly integrated into police department workflows through vendors like Clearview AI, allows investigators to run images against databases of hundreds of millions of faces scraped from social media without consent.
Then there are the Smart City contracts — broad municipal agreements with technology companies like Palantir, Genetec, and various Amazon Web Services subsidiaries — that promise to integrate all of these data streams into unified analytics platforms. Cities are sold these systems with language borrowed from Silicon Valley: efficiency, optimization, predictive capability. What they are actually purchasing is the capacity to monitor the movement, behavior, and associations of their residents at a scale that would have been unimaginable to any previous generation of civil libertarians.
Where the Cameras Point
The distribution of this surveillance infrastructure is not random, and the evidence that it concentrates in communities of color is not anecdotal. A 2021 report by the Surveillance Technology Oversight Project analyzed camera placement data in New York City and found that surveillance density was significantly higher in predominantly Black and Latino neighborhoods than in comparable white neighborhoods with similar or higher crime rates. A separate analysis of Chicago's network of more than 32,000 cameras found analogous patterns — monitoring concentrated in South and West Side neighborhoods, with far sparser coverage on the North Side.
This geographic concentration is not merely an equity concern in the abstract. It has direct and documented consequences for who gets stopped, investigated, and arrested. Algorithmic policing tools — systems that use historical crime data to predict where future crimes will occur and who is likely to commit them — are trained on data generated by prior policing patterns. When those patterns already reflect racial bias, the algorithm learns and amplifies that bias. A neighborhood that has historically been over-policed generates more crime data, which the algorithm interprets as evidence of higher criminality, which justifies more policing, which generates more data. The feedback loop is self-confirming and, from inside the system, invisible.
Facial recognition technology compounds the problem. The National Institute of Standards and Technology has published repeated findings that leading facial recognition algorithms produce significantly higher error rates for Black women, dark-skinned individuals, and people of East Asian descent than for white men. At least three documented cases — Robert Williams in Detroit, Michael Oliver in New Orleans, Nijeer Parks in New Jersey — involved Black men wrongfully arrested based on facial recognition misidentification. In each case, the technology identified a suspect. In each case, the technology was wrong. In each case, a man was handcuffed, jailed, and forced to prove his innocence.
The Democratic Deficit
Perhaps the most striking feature of this surveillance expansion is how little of it has been subject to meaningful democratic deliberation. Contracts between police departments and surveillance vendors are routinely classified as law enforcement sensitive, exempting them from public records requests. City councils are presented with technology acquisitions as line items in equipment budgets rather than as policy decisions with civil liberties implications. When community groups have sought to review the terms of Smart City contracts, they have frequently encountered non-disclosure agreements that prohibit the city itself from revealing what data is collected, how long it is retained, and who can access it.
Amazon's Ring partnership agreements with police departments, obtained through Freedom of Information Act requests by journalists and advocacy groups, revealed that the company placed no restrictions on how police could use footage obtained through the portal, required no warrant or judicial authorization for most requests, and reserved the right to provide footage directly to law enforcement without the homeowner's consent in circumstances it defined unilaterally.
The public safety argument made by proponents of these systems deserves a fair hearing before it is challenged. The claim is straightforward: cameras deter crime, footage solves crimes, and the communities that benefit most from crime reduction are often the same communities that bear the highest burden of violent crime. This argument has genuine moral weight. No serious progressive commentator should dismiss the real and disproportionate harm that violent crime inflicts on low-income communities of color.
But the evidence that mass surveillance actually delivers on its public safety promise is far weaker than its advocates suggest. A comprehensive 2019 study published in the Journal of Experimental Criminology found that CCTV camera systems produced modest and inconsistent effects on crime rates, with the strongest effects limited to parking lots and other confined environments. Chicago's own evaluation of its camera network found no statistically significant reduction in violent crime in areas with high camera density. Meanwhile, the social costs — chilling effects on protest and assembly, documented misidentifications, the normalization of monitoring as a condition of public life — are borne entirely by the communities being watched.
What Accountability Would Look Like
A growing number of cities have begun to push back. San Francisco, Oakland, and Boston have enacted bans or significant restrictions on government use of facial recognition technology. The American Civil Liberties Union and allied organizations have advanced model legislation requiring city council approval for any surveillance technology acquisition, mandatory civil liberties impact assessments, and public reporting on how surveillance tools are used and who they are used against.
These are not radical proposals. They are the minimum conditions for democratic governance of technologies that have the capacity to reshape the relationship between the state and the citizen. The question of whether a city should build a comprehensive surveillance network is a political question — a question about values, about the kind of community residents want to live in, about who bears the costs of security and who receives its benefits. It is not a procurement question to be resolved by a police chief and a technology vendor in a conference room.
When corporations like Amazon and Google sell surveillance infrastructure to governments, they are not merely selling hardware and software. They are inserting themselves as permanent, profit-motivated intermediaries in the relationship between the state and the people it is supposed to serve. The data generated by these systems is valuable — commercially valuable — in ways that extend far beyond any law enforcement application. And the communities whose movements, behaviors, and associations are being logged, analyzed, and retained indefinitely have, in most cases, never been asked whether they consent.
A democracy that allows its most vulnerable communities to be placed under permanent corporate-assisted surveillance — without their knowledge, without their consent, and without any meaningful mechanism for accountability — is not governing in the public interest. It is renting out its citizens to the highest bidder, one municipal contract at a time.
