Kashmir Hill (Simon & Schuster)
IN 2011, two things happened within a few weeks. Eric Schmidt, then executive chair of Google, said the company had decided not to build a facial recognition database because doing so was 鈥渃rossing the creepy line鈥�. Then Facebook released just such a feature, automatically tagging people in uploaded photographs, grouping them and asking users to identify anyone it couldn鈥檛. It eventually discontinued the system in 2021.
The creepiness has spread. Police in the UK used live facial recognition to scan for criminals at the coronation of King Charles. In New York, Madison Square Garden Entertainment using the tech to bar from its venues lawyers working for firms involved in litigation against it. The world鈥檚 airports are this tech in new infrastructure. In China, facial recognition has been to stop theft. In the US, false matches have led to false arrests (and lawsuits).
Advertisement
These are all the kinds of things Kashmir Hill, a reporter for The New York Times, imagined when she first learned of Clearview AI via a tip-off. The tech firm had amassed a database of billions of faces scraped off the web, and its business model rested on selling access to that information to law enforcement. Having long specalised in privacy issues, Hill understood the size of the threat that the database posed.
Almost immediately, she found the firm was so secretive that every contact stopped answering her emails after her first questions. But Hill persisted, following leads and developing sources. In a January 2020 , she alerted the world to 鈥渢he secretive company that might end privacy as we know it鈥�.
For the first half of her highly readable new book, Your Face Belongs to Us: The secretive startup dismantling your privacy, Hill鈥檚 chapters alternate between the origins of Clearview and the general history of facial recognition. These twin threads eventually join into a single narrative about the present and future of this technology.
Key among her characters is Clearview CEO Hoan Ton-That, who emigrated to San Francisco from Australia at 19 and spent a decade experimenting with a variety of uninspired apps (including, inevitably, one to let you rate other people鈥檚 appearance) before moving to New York and hitting on the idea of face matching. As Hill makes clear, the technological pieces were already within reach. Crossing the creepy line just took an ambitious entrepreneur with no inconvenient ethical concerns, steeped in the Silicon Valley ethos of 鈥渁sk forgiveness, not permission鈥�.
Finding investors was expedited by the 6 January 2021 insurrection, when supporters of Donald Trump stormed the US Capitol building in Washington DC. Clearview helped investigators seeking to identify the participants 鈥� apparently unhindered by mounting in the or by its business-limiting with the state of Illinois under the Biometric Information Privacy Act.
As Hill鈥檚 later chapters show, if we do nothing, the de-anonymised future hurtling towards us will allow anyone with a modicum of technical ability to cobble together a face search engine to fit in their pocket. And not just face: besides the potential for DNA samples and other biometrics, enough video is posted daily to enable a database of highly identifiable gaits.
Towards the end of the book, we find Ton-That experimenting with a pair of glasses that can isolate any face in your field of vision and show matches from a Clearview search. In the world he demonstrates with those glasses, one that Hill fears, no one will ever be anonymous in public again 鈥� a liberty we may only fully understand when it is gone.
Wendy M. Grossman is a technology writer based in London
