Weibo user claims to have matched 100K women from porn with facial recognition
A Germany-based user of Chinese social media network Weibo says he has used biometric facial recognition to match more than 100,000 women in adult videos to social media profiles, sparking privacy concerns. A post to Weibo went viral and was reported by Motherboard, and has also been translated and posted to Twitter by a Stanford political science PhD candidate, where it went viral again.
The individual provided no proof of the claim, and when contacted by Motherboard said that “database schema” and technical details will be provided next week. The individual and some programmer friends scanned videos to enable men to find out if their relationship partner has participated in online pornography, according to initial claims, or to enable women to detect copyright infringement, according to subsequent claims.
“This is horrendous and a pitch-perfect example of how these systems, globally, enable male dominance,” tweeted Soraya Chemaly, author of Rage Becomes Her, as reported by Motherboard.
In 2016, reports out of Russia indicated that people were using the FindFace facial recognition service to identify and intimidate women who had appeared in pornography or advertised on prostitution and escort service websites. Fake pornography was also one of the concerns expressed by researchers about deepfake technology last year.
The latest attempt to break other people’s anonymity measures comes at a time when controversy around facial recognition has become mainstream consumer news, between widespread surveillance and alleged repression in Xinjiang, a ban on the technology in San Francisco, and U.S. Congress slowly wading into questions around appropriate use by police and potential regulation.
Comments