AI found misused to aid porn
Face-changing tech urged to be strictly controlled
The fast development of artificial intelligence (AI) technology has opened an underground market for pornography with the face of female celebrities, which has intensified concerns over privacy.
An underground market of fake porn with celebrities’ faces has emerged in China with the reduced costs of AI technology that can replace an actor’s face with another person’s.
The Global Times on Thursday found that underground sellers of porn videos with female actresses’ faces used argot to attract buyers at online forums and share a cloud drive link of the videos after receiving money through Wechat or Alipay.
The price depends on the length and quality of the video, and a package of 700 videos is sold for 158 yuan ($23), The Beijing News reported Thursday.
Customers can also tailor their videos on Xianyu, an online marketplace of used goods. A service provider charged 40 yuan per minute of the video and offered to transact on Wechat when approached by the Global Times posing as a client.
“The more photos you provide, the more natural the video will be. A one-minute video takes a few hours to make,” said the seller based in Bozhou, East China’s Anhui Province.
The seller also offers a bundle of face-changing application “deepfake” and tutoring for 400 yuan. The technology was previously used to splice images of Hollywood A-listers, such as Wonder Woman actress Gal Gadot and Harry Potter star Emma Watson into pornographic content.
The fake porn industry has intensified public concerns over privacy and information security, with many saying that the technology could “make porn of people you hate.”
Qin An, head of the Beijing-based Institute of China Cyberspace Strategy, noted that the development of technology is ahead of legislation and China has no law on AI yet.
“Producers of the underground porn chain, within the existing legal framework, would be punished for disseminating pornographic articles and violating the image rights of others,” Qin said.
Beijing-based technology company MEGVII, which specializes in facial recognition, issued a code of conduct on Wednesday which said the use of AI must be supervised, traceable and prudent.
AI solutions should be verifiable and traceable, and all the mistakes and faults must be addressed once they are spotted, the code said.
“AI service providers shall include a privacy regulation in its terms of service,” a spokesperson of MEGVII told the Global Times on Thursday.
Industry regulation and a law guarantee the industry’s healthy and sustainable development, said the spokesperson.
In July 2017, China issued a plan on new generation AI, which will make it a major new growth engine and improve people’s lives. It also set the goal of China becoming a major center and world leader in AI innovation by 2030. Chen Qingqing contributed to this story