Data protection regulators in Italy have also opened an investigation into its use, and access to the bot has been restricted on Apple’s iOS.The scrutiny of Telegram comes as multiple investigations into the use of the messaging service have also discovered private groups sharing non-consensual “revenge porn” photos and videos of women that were not generated by AI.
Sensity was only able to count images shared publicly and the bot gives people the option to generate photos privately.The images are automatically created once people upload a clothed image of the victim to the Telegram bot from their phone or desktop.
Deepfake videos hosted on three of the biggest porn websites, XVideos, Xnxx, and xHamster, have been viewed millions of times.Deepfake upload figures seen by WIRED did not include Pornhub, which is the second-biggest porn website and despite banning deepfakes in 2018 still has problems with the videos.
To make them, Tully needed only to gather a few hundred images of Hanks online and spend less than $100 to tune open-source face-generation software to his chosen subject.
“It’s all fine and good for helping human moderators, but it's obviously not even close to the level of accuracy that you need,” says Hany Farid, a professor at UC Berkeley and an authority on digital forensics, who is familiar with the Facebook-led project.
That day, representatives learned that a “high school kid with a good graphics card can make this stuff.” That the creators of malicious deepfakes (the bad guys) and those working to identify and intercept fake content (the good guys) are locked in an unending arms race.
Jerome Pesenti: Moderating automatically, or even with humans and computers working together, at the scale of Facebook is a super challenging problem.Recently there’s been a lot of progress in the field of language , allowing us a much more refined understanding of interactions through the language that people use.