What happens when surveillance gets it right?
Public resistance to Flock raises a deeper question: are we beginning to worry less about what happens when surveillance identifies the wrong person, and more about what happens when it identifies the right one?
A year ago this month, the New York Times published my visual op-ed, “The N.Y.P.D. Is Teaching America How to Track Everyone Every Day Forever.”
The last decade of my life has been spent working on the issues that piece tried to make visible. I began focusing on surveillance technology as a line public defender in Washington, D.C. Then, I built a project at Brooklyn Defender Services around science, surveillance, and the criminal legal system. And I soon slipped from direct representation of the accused to targeting a significant part of my work toward something much harder to measure than a case outcome: narrative change.
So one year later, what interests me is not that my analysis of surveillance has changed. It hasn't. It is the possibility that the public’s relationship to it has.
The Times piece mapped a surveillance infrastructure New York had built over decades: cameras, license plate readers, facial recognition, social media monitoring, location information, police databases, and other systems that increasingly allow government to reconstruct where people go, whom they associate with, and what they do. The argument was never simply that some of those technologies might be inaccurate or abused. It was that, taken together, they fundamentally change the relationship between people and the state.
One of the hardest editorial fights over that piece was keeping even some of that larger argument intact. I fought particularly hard to preserve Justice Louis Brandeis’s description of the “right to be let alone.” In his Olmstead dissent nearly a century ago, Brandeis connected that right not merely to secrecy, but to the protection of our thoughts and beliefs against government intrusion.
But the word I most wanted to use was largely cut from the final piece.
Anonymity.
For years, I have believed that anonymity is one of the conditions that makes democratic life possible.
I do not mean a generalized constitutional right to move everywhere without being identified. The law is more complicated than that. But the Supreme Court has recognized the importance of anonymity in political life. In McIntyre v. Ohio Elections Commission, a case involving anonymous political speech, the Court called anonymity “a shield from the tyranny of the majority,” emphasizing its protection of unpopular people and ideas.
The principle reaches beyond anonymous political pamphlets.
Anonymity gives us room to move through the world without every action automatically attaching itself to a permanent, named record. It creates space to associate, explore, dissent, change our minds, try on identities, encounter unpopular ideas, and participate in public life without always calculating how an institution with power over us might interpret those choices later.
And for much of modern life, we possessed a substantial degree of practical anonymity even when we were technically observable.
You could attend a rally in Washington Square Park. Ride across town. Visit a doctor. Walk into a place of worship. Meet someone for coffee. Spend an afternoon with people the police considered suspicious.
Someone might see you.
But following you required something.
Police had to decide to watch you. Someone had to plan the surveillance, assign personnel, spend money, maintain attention, and continue doing it over time. Human limitations created practical barriers between being visible in public and being persistently knowable to the state.
Modern surveillance systems erode those barriers.
That was one of the ideas I tried hardest to convey in the Times piece. When data can be collected first and searched later, the government’s attention does not always have to fall on you in advance. The record can already exist. The reason for becoming interested in you can come later.
The glare of state attention can swivel backward in time.
For years, I talked about this in rooms full of public defenders, civil rights lawyers, advocates, organizers, and communities. And one of the hardest narrative problems we identified was remarkably persistent.
Surveillance keeps us safe.
The cameras help police find dangerous people. The databases help solve crimes. More information gives law enforcement more tools. And if you have done nothing wrong, greater visibility to government is largely someone else’s problem.
That story, of course, was never shared equally. Communities subjected to intensive policing have long understood that being accurately identified, categorized, located, and tracked by the state can itself create danger. Simone Browne's Dark Matters traces that history much further back, showing how surveillance, racism, and the monitoring of Black life have long been intertwined. But as a broader public narrative, the connection between surveillance and safety proved remarkably difficult to disturb.
In our daily lives, surveillance had also became ordinary one concession at a time. A camera here. A database there. Longer retention. More sharing. Another source of information becoming searchable. Woodrow Hartzog, Johanna Gunawan, and Evan Selinger called these incremental erosions "privacy nicks.” Selinger and Divya Ramjee recently went on to connect that idea to the backlash against Flock, arguing that one reason this particular technology has broken through is salience. Its physical presence makes the network itself easier to see. (Tech Policy Press)
That is a persuasive explanation.
But after years spent trying to move this conversation, I keep wondering whether something else is happening too.
Facial recognition offered an earlier moment of broad public alarm. Much of the public-facing coverage that propelled it into wider consciousness centered on failure: false matches, wrongful arrests, racial disparities. Those harms are real and enormously important. But they also made the problem with surveillance relatively easy to understand: the system identified the wrong person.
Critics were simultaneously making another argument: a system capable of identifying people cheaply, ubiquitously, and accurately poses its own danger. But that argument is harder to make visceral. What is the story when surveillance identifies exactly the person the government intended to find?
Flock makes me return to that narrative challenge.
Over the last year, public attitudes toward automated license plate readers have flipped. A recent Washington Post article reported, "A YouGov survey this week of approximately 20,000 U.S. adults found that 46 percent said they oppose the technology being used in their communities, compared with 38 percent support. When asked the same question last year, 33 percent voiced opposition and 45 percent voiced support." Flock has challenged the survey’s framing and argues that support rises when people are told more about the technology’s investigative uses. (The Washington Post)
But communities are also already acting. Across the country, localities have paused deployments or walked away from Flock, with resistance emerging across conventional political lines. (Axios) (ABC)
I do not have a satisfying theory for why this particular technology, at this particular moment, has activated people across those lines. Selinger and Ramjee may be right that visibility is central. Distrust of institutions may matter. Immigration enforcement may have changed how some people understand what it means to be findable. Police misuse certainly gives people concrete reasons for concern. Perhaps years of small surveillance incursions simply accumulated until people could finally see the system.
Probably no single explanation is sufficient.
But I keep coming back to a simpler question.
Are we moving from a public debate dominated by what happens when surveillance identifies the wrong person toward one increasingly concerned with what happens when surveillance identifies the right one?
What if the camera sees you correctly?
What if the system accurately records where your car was yesterday, last month, or six months ago?
What if those records can be searched after the reason for searching has changed?
What if the person querying the system is looking for exactly the person they intend to find?
None of those questions depends on technological error. Accuracy does not resolve them. In some circumstances, the problem is precisely that the systems are accurate, persistent, and integrated.
Flock makes that easier to see. A license plate reader sounds like a discrete thing: a camera photographs a passing vehicle and records its plate and other vehicle details. But that description misses almost everything important. That photo becomes a time-and-location record. Records accumulate. They become searchable. Cameras become networks. Networks extend across jurisdictions. Ordinary movement becomes historical data that can be queried after the fact.
So the camera is a technology. But Flock is a system. And it is a symptom of something larger. I have come to describe that larger infrastructure as the surveillance stack: accumulated layers of images, location trails, communications, biometric information, police records, commercial data, and other traces that can increasingly be searched, analyzed, and connected.
What matters about a stack is not simply that each layer exists. It is the capacity that emerges when the layers can work together.
Perhaps that is another reason this moment feels different. For years, surveillance infrastructure was remarkably easy to experience one component at a time. The cumulative system was harder to see. In some ways, that was the entire project of my Times piece: to put those technologies beside one another and ask readers to see the infrastructure they created together. I fought to keep Brandeis’s “right to be let alone” in the piece because the stakes were never merely technical.
What disappears as persistent surveillance becomes ordinary is not simply secrecy. It is anonymity.
That is why I want to be careful about declaring the current backlash a victory over the surveillance state. People can reject Flock without rejecting the underlying capacity. A city could terminate its contract with Flock and execute another with Axon. (The Guardian) The cameras could change while the data flows remain. Flock could become, like facial recognition before it, the frightening technology we single out while leaving the larger architecture intact.
Nor do I want to suggest that Americans have only now discovered that government visibility can be dangerous. Many communities have been living that reality for generations.
What interests me is the possibility that an intuition long familiar to heavily surveilled communities is moving into a broader public consciousness:
Being accurately seen by the state is not always protection. Sometimes it is exposure.
That is different from the familiar privacy debate over whether innocent people have anything to hide. You may have nothing to hide. You may simply need somewhere to be anonymous.
For nearly a decade, I have been trying to persuade rooms of lawyers and policymakers that the loss of anonymity was itself a democratic harm. That making people persistently identifiable, searchable, and retrievable changes something fundamental even when the technology works exactly as designed.
A year ago, I was still making that argument in the Times.
The argument has not changed.
What may be changing is the audience’s relationship to it.
I do not know whether the backlash against Flock means the old equation of surveillance with safety is actually breaking down, or whether this moment will survive the next frightening crime story, the next technology, or the next promise that a little more visibility will finally make us secure.
But after spending a decade trying to move this narrative, the questions themselves sound different to me.
Maybe more people are beginning to understand anonymity not as the refuge of someone with something to hide, but as something ordinary people need in order to live free.
If so, that would be a much bigger change than the rejection of any one surveillance technology.
Because the question underneath Flock is not really whether we want cameras on poles.
It is whether we are willing to build a society in which there is nowhere left to be anonymous.