 We’re not sure which is most disconcerting: that Meta has a division named Global Threat Disruption, that their idea of said global threats includes deepfake celebrity endorsements, or that this has become their excuse to reactivate the controversial facial recognition software they shelved just three years earlier (so much for the “Delete” key). Meta has relaunched DeepFace to defend against celebrity deepfakes in South Korea, Britain, and even the European Union. “Celeb-baiting,” as it’s known, is where scammers populate their social media posts with images or AI-generated video of public figures. Convinced that they’re real – that Whoopi Goldberg really is endorsing a revolutionary weight loss system, for example – unwitting victims fork over their data and money with just a few clicks. All of which, according to Meta, “is bad for people that use our products.” Celeb-baiting is a legitimate problem, to be sure. We’re no fans of social media scammers. What’s more, we know full well that “buyer beware” is meaningless in a world where it is increasingly difficult to spot digital fakes. But in reviving their facial recognition software, Meta may be rolling out a cannon to kill a mosquito. The potential for collateral damage inherent in this move is, in a word, staggering. Just ask the Uighurs in Xi’s China. Meta began tracking the faces of one billion users, beginning in 2015. And initially, it didn’t bother to tell people the technology was active, so users couldn’t opt out. As a result of Meta’s sleight of hand, as well as its own strict privacy laws, the EU cried foul and banned DeepFace from being implemented. But that was years ago … and how times have changed. The privacy-minded Europeans are now letting Meta test DeepFace to help public figures guard against their likenesses being misused. But can regular users be far behind? Meta could rebuild its billion-face database in no time. For its part, the U.K. is courting artificial intelligence like never before, declaring that it will help unleash a “decade of national renewal.” Even for a country that never met a facial recognition system it didn’t love, this feels like a bridge too far. We have written about the dangers, both real and looming, of a world in which facial recognition technology has become ubiquitous. When DeepFace was shelved in 2021, it represented an almost unheard-of reversal, in effect putting the genie (Mark Z, not Jafar) back in the bottle. That incredibly lucky bit of history is unlikely to repeat itself. Genies never go back in their bottles a second time. Comments are closed.
|
Categories
All
|
RSS Feed