Bella carrellata sull'uso delle tecnologie biometriche per law enforcement, dal "bertillonage" al riconoscimento facciale, inclusi abusi e errori. <https://slate.com/technology/2018/03/with-apples-face-id-its-time-to-look-at...> At Apple’s near-sacred product unveiling event last year, the iPhone X was undoubtedly the star of the show. Among its most boasted about features? The sleek device’s new security recognition system, Face ID. Rather than asking users to use their fingerprint on the now-nonexistent home button to unlock their phones, the iPhone X’s Face ID uses its cameras to make 3-D scans of their faces, which then enable them to unlock their phones by just holding the device up to their mugs. At the event, an exec boasted <https://www.wired.com/story/iphone-x-faceid-security/> that the facial recognition technology proved far safer than its previous fingerprint-based Touch ID, claiming that there’s only a 1 in a million chance of a random stranger’s face unlocking a user’s device. Enthusiastic users across the globe recorded themselves <https://www.youtube.com/results?search_query=testing+iphone+X+FaceID> attempting to game the new feature. Nearly all failed: elaborate masks, lightning variations, makeup, costumes, and even twins tried to cheat Face ID, mostly without success. Faces have emerged in our imaginations as a sort of walking passcode of human nature, waiting to be deciphered and explained. The hype over Face ID’s futuristic promises of security with ease, however, may be undercut by more problematic recent applications of facial recognition technology. In September, for example, two Stanford researchers released a study <https://osf.io/zn79k/> that dubiously claimed <http://www.slate.com/blogs/future_tense/2017/09/14/be_suspicious_of_that_stu...> that a machine-learning system predicted, with 74 to 81 percent accuracy, whether a woman or man was gay or straight based solely on photos of their faces. That same month, a team of researchers from Cambridge University, India’s National Institute of Technology*//*Warangal, and the Indian Institute of Science presented a paper <https://arxiv.org/pdf/1708.09317.pdf> on a machine-learning system that they said could identify protesters hidden behind simple caps and scarves with success rates of about 69 percent. Both studies had their flaws but raised legitimate concerns about how advancing facial recognition technologies could be used to profile, intimidate, harass, or abuse vulnerable populations like minorities, political dissidents, and the poor. We shouldn’t be surprised. As revolutionary as these capabilities may seem, the idea of analyzing faces with technology to confirm identities, make judgments, or predict behavior is nothing new. Instead, it seems faces are just back in vogue, and with a vengeance. Though they may have fallen out of favor in recent decades (dismissed as racist pseudoscience, as was the case with physiognomy <http://blogs.getty.edu/iris/physiognomy-the-beautiful-pseudoscience/>, or edged out with other biometrics, such as fingerprints and DNA), over the past several centuries, humans have invested extraordinary amounts of resources trying to detect patterns in one another’s countenances that we think will reveal something about the people behind them. Faces have emerged in our imaginations as a sort of walking passcode of human nature <https://books.google.com/books?id=oQOcCwAAQBAJ&pg=PT206&lpg=PT206&dq=%22%E2%...>, waiting to be deciphered and explained. Creators of today’s facial recognition and analysis tools claim that their technologies are different. Powered by high-tech cameras and advanced algorithms, they hold out the promise of precision and scientific objectivity. But the past provides cautionary examples of the limits of such projections—showing how our conception of accuracy has changed, and how our use and misuse of these identification tools can come at great social costs. [...]