Body cameras and the police-state concern
In a discussion of police body cameras, Matthew Feeney raised facial recognition as the danger that proper rules must forestall. If body camera footage were never deleted, he said, it would not be hard for a nefarious local government to track citizens through that footage by implementing facial recognition software; body cameras could then quickly become a device of a police state Matthew Feeney on Best Practices for Police Body Cameras (2016). He framed the question in general terms, asking what a totalitarian would think of a given law enforcement technology, and grouping body cameras with drones and militarized equipment as attractive to people who do not care about individual liberty — while holding that with the right rules they can be made tools of transparency and accountability.
Bob Zadek expanded on the point: technology gives government so much power that citizens depend to some degree on government not misusing it. A totalitarian state and a free state would have the same technology, he said, and all that separates them is how it is used; voters are empowering whoever they elect to be at the on-and-off switch Matthew Feeney on Best Practices for Police Body Cameras (2016). Feeney agreed, adding that he is impressed by such technology only because of laws, and that in a different context he would find it terrifying.
Anonymity, private data and the limits of constitutional law
Jeff Kosseff, discussing anonymity in America, located the principal threat to anonymous speech not in government subpoenas or new laws but in private companies’ growing control over personal data and speech The United States of Anonymous (2022). He described facial recognition and geolocation as data that brokers can sell to anyone in the public, and as information that can trace people back to their online or offline speech. He cited Kashmir Hill’s reporting in the New York Times on Clearview AI, describing it as a technology used by local police — he said more than 600 police departments — produced by a company that scrapes social media pictures and can automatically identify someone by facial recognition. That, he said, poses the greatest challenge, and requires privacy laws that grapple meaningfully with private data.
Zadek pressed an analogy: facial recognition, he suggested, is nothing other than a private actor seeing you in the street and recognizing you, only on steroids; a recognizable celebrity or politician already has less privacy because many people know who they are The United States of Anonymous (2022). Kosseff answered that it depends on how the data is gathered, that it is sometimes merged with other private information maintained about a person, and that from a Fourth Amendment perspective — though post-Carpenter he was not willing to say how that shakes out — information in public view is going to be hard to control, including data the government uses that you have publicly posted.
Kosseff argued the response should come through privacy law rather than the First or Fourth Amendment. He noted that California has passed two laws giving people the ability to access data stored about them and to request its deletion, calling that a positive step but not a panacea, partly because he does not know every data broker that holds information about him and so would not know whom to ask. He proposed that law could specify the ways facial recognition can and cannot be used, and said the United States has never had a meaningful national privacy law — that plenty have been proposed, but every interest group seeks carve-outs, and Hill staffers and members he speaks with treat passage as very difficult. He contrasted American treatment of privacy with Europe’s, where privacy is viewed as a fundamental human right, and cited Europe’s right to be forgotten as something he did not think could exist here.
Facial recognition among other surveillance tools
Jonathan Hofer, asked how much of a discussion of license plate readers applies with equal force to facial recognition, said the issues are very similar: the tools are used similarly tactically by police and share the same data collection and step-tracing problems The Mass Surveillance nobody is talking about (2022). He judged license plate readers the bigger concern based on their number, saying facial recognition is far and few between and that a lot of cities have outright banned it — naming Seattle, Portland, San Francisco and a few others on the East Coast — while most cities have not banned license plate readers and they are used everywhere.
Hofer said one of his big fears is that the United States will quickly adopt tactics and strategies the CCP uses to control its Uyghur population, noting that the Chinese government has used surveillance tools, most notably facial recognition, and maintains large databases of purchases and internet search history. He added that the United States already has internet search history on lockdown with the NSA, and that few people realize the United States is one of the largest importers of Chinese surveillance technology in the world. He nonetheless saw reason for optimism, pointing to votes that have passed in the Bay Area and elsewhere curbing government encroachment on privacy, which he thought could be replicated.
Zadek’s closing position was that the technology will not be stopped, so the issue is not whether government has or uses technology to collect but the use: misuse should bring jail or a fine, and the aim should be limiting what government is allowed to do with the technology rather than stifling or controlling the technology itself The Mass Surveillance nobody is talking about (2022).
Across episodes: regulating use, not technology
The same question — whether the danger lies in the tool or in its use — runs through all three episodes, and the treatment shifts in emphasis rather than in conclusion. Feeney in 2016 treats facial recognition as a hypothetical add-on to retained body camera footage and answers with proper rules and regulation; Kosseff in 2022 treats it as an existing private-sector capability, embodied in Clearview AI and data brokers, and answers with national privacy law; Hofer in 2022 treats it as one surveillance tool among others, already banned in several cities, and answers with local votes and optimism. Zadek advances the use-not-technology formulation in the Feeney episode and states it most fully in the Hofer episode, where he says the prospect of controlling use is difficult but there is reason to be optimistic, while there is no prospect of controlling the technology.
What the sources do not cover
The excerpts do not describe any facial recognition statute, case or amendment holding in detail, and no bill is named. They do not state how any city’s ban was enacted or what it prohibits, nor how Clearview AI’s scraped database is assembled beyond the fact of scraping social media pictures. Kosseff’s reference to Carpenter is left as an unresolved question about how the law shakes out, and the excerpts do not say what that case held.