This Friday (September 22), I'm (virtually) presenting at the DIG Festival in Modena, Italy. That night, I'll be in person at LA's Book Soup for the launch of Justin C Key's "The World Wasn’t Ready for You." On September 27, I'll be at Chevalier's Books in Los Angeles with Brian Merchant for a joint launch for my new book The Internet Con and his new book, Blood in the Machine.
Your Face Belongs To Us is Kashmir Hill's new tell-all history of Clearview AI, the creepy facial recognition company whose origins are mired in far-right politics, off-the-books police misconduct, sales to authoritarian states and sleazy one-percenter one-upmanship:
Hill is a fitting chronicler here. Clearview first rose to prominence – or, rather, notoriety – with the publication of her 2020 expose on the company, which had scraped more than a billion facial images from the web, and then started secretly marketing a search engine for faces to cops, spooks, private security firms, and, eventually, repressive governments:
Hill's original blockbuster expose was followed by an in-depth magazine feature and then a string more articles, which revealed the company's origins in white nationalist movements, and the mercurial jourey of its founder, Hoan Ton-That:
The story of Clearview's technology is an interesting one, a story about the machine learning gold-rush where modestly talented technologists who could lay hands on sufficient data could throw it together with off-the-shelf algorithms and do things that had previously been considered impossible. While Clearview has plenty of competitors today, as recently as a couple of years ago, it played like a magic trick.
That's where the more interesting story of Clearview's founding comes in. Hill is a meticulous researcher and had the benefit of a disaffected – and excommunicated – Clearview co-founder, who provided her with masses of internal communications. She also benefited from the court documents from the flurry of lawsuits that Clearview prompted.
What emerges from these primary sources – including multiple interviews with Ton-That – is a story about a move-fast-and-break-things company at the tail end of the forgiveness-not-permission era of technological development. Clearview's founders are violating laws and norms, they're short on cash, and they're racing across the river on the backs of alligators, hoping to reach the riches on the opposite bank without losing a leg.
A decade ago, they might have played as heroes. Today, they're just grifters – bullshitters faking it until they make it, lying to Hill (and getting caught out), and the rest of us. The founders themselves are erratic weirdos, and not the fun kind of weirdos, either. Ton-That – who emigrated to Silicon Valley from Australia as a teenager, seeking a techie's fortune – comes across as a bro-addled dimbulb who threw his lot in with white nationalists, MAGA Republicans, Rudy Guiliani bagmen, Peter Theil, and assorted other tech-adjascent goblins.
Meanwhile, biometrics generally – and facial recognition specifically – is a discipline with a long and sordid history, inextricably entwined with phrenology and eugenics, as Hill describes in a series of interstitial chapters that recount historical attempts to indentify the facial features that correspond with criminality and low intelligence.
These interstitials are woven into a-ha moments from Clearview's history, in which various investors, employees, hangers-on, competitors and customers speculate about how a facial-recognition system could eventually not just recognize criminals, but predict criminality. It's a potent reminder of the AI industry's many overlaps with "race-science" and other quack beliefs.
Hill also describes how Clearview and its competitors' recklessness and arrogance created the openings for shrewd civil libertarians to secure bipartisan support for biometric privacy laws, most notably Illinois' best-of-breed Biometric Information Privacy Act:
But by the end of the book, Hill makes the case that Ton-That and his competitors have gotten away with it. Facial recognition is now so easy to build that – she says – we're unlikely to abolish it, despite all the many horrifying ways that FR could fuck up our societies. It's a sobering conclusion, and while Hill holds out some hope for curbing the official use of FR, she seems resigned to a future in which – for example – creepy guys covertly snap photos of women on the street, use those pictures to figure out their names and addresses, and then stalk and harass them.
If she's right, this is Ton-That's true legacy, and the legacy of the funders who handed him millions to spend building this. Perhaps someone else would have stepped into that sweaty, reckless-grifter-shaped hole if Ton-That hadn't been there to fill it, but in our timeline, we can say that Ton-That was the bumbler who helped destroy something precious.
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
With bans on phones and social media, we are disarming students who refuse to swallow the bunting.
Blanket phone bans are a failure to do holistic, systems thinking and adopt structural ideology. Those require community conversation. Those require understanding and recognizing ecosystems. Those require thinking in terms of prevention (in social work terms) instead of tier 1 intervention (in RTI terms).
Phone bans are bikeshedding.
Phone bans are fundamental attribution error.
Phone bans are…
Having a client that used to be a cop is so... not enlightening really but like... just a little peek into the particular brand of busybody who signs up to be a cop because literally anything happens in the house and they're knocking on the office door asking who what where when like girliepop yall heard of client confidentiality I ain't telling you shit come back with a warrant
Students have to download software prior to the exam. At 8:30 a.m. Tuesday, they have to open the download. The software will then disconnect the Internet on the laptop and turn on the computer's camera and microphone. Students must be in alone in the room — any noise, from children to a pet, could flag the video for cheating. When they log into the system, students must remain still in front of the camera so it can take a picture of them and compare it with a previously uploaded photo, according to the directions issued to test takers.
Some of those taking the exam are questioning the move online and worry about glitches and tracking software.
Before we build, we need to dismantle the surveillance ed-tech that already permeates our schools. And we need to dismantle the surveillance culture that it’s emerged from. I think this is one of our most important challenges in the months and years ahead. We must abolish “cop shit,” recognizing that almost all of ed-tech is precisely that.
Cop shit is seductive. It makes metrics transparent. It allows for the clear progress toward learning objectives. (“Badges” are cop shit, by the way.) It also subsumes education within a market logic. “Here,” cop shit says, “you will learn how to do this thing. We will know you learned it by the acquisition of this gold star. But in order for me to award you this gold star, I must parse you, sense you, track you, collect you, and—” here’s the key, “I will presume that you will attempt to flout me at every turn. We are both scamming each other, you and I, and I intend to win.” When a classroom becomes adversarial, of course, as cop shit presumes, then there must be a clear winner and loser. The student’s education then becomes not a victory for their own self-improvement or -enrichment, but rather that the teacher conquered the student’s presumed inherent laziness, shiftiness, etc. to instill some kernel of a lesson.
No wonder the traditional humanities classroom of “read things, come together and talk about them, and write papers about them” has disappeared in the age of cop shit. There’s no game to fix, no battle to win.
