China considers legislating face, voice biometric data rules for deepfakes
The Cyberspace Administration of China (CAC) has unveiled draft rules intended to regulate the production and distribution of fakes such as the manipulation of face or voice biometric data using deep learning or virtual reality technologies, Reuters reports.
Per the proposed rules which have been presented to the public for comment, deepfake content providers will henceforth be required to respect what the CAC refers to as “social morality and ethics,” as well as “the correct political direction.”
The rules also require content creators to obtain consent before someone’s text, audio, images or videos can be impersonated or used for disinformation or manipulation purposes.
While the rules provide for a system where those offended can channel their complaints, they also suggest fines of between 10,000 and 100,000 yuan (roughly US$1,600 and $16,000) for first-time offenders, but rule breaches could also attract criminal sanctions, according to Reuters. Under the proposed legislation, app stores would also be called upon to remove providers of deepfake content technologies from their platforms.
Reuters quotes Article 12 of the draft legislation which reads: “Where a deep synthesis service provider provides significant editing functions for biometric information such as face and human voice, it shall prompt the (provider) to notify and obtain the individual consent of the subject whose personal information is being edited.”
The CAC also highlights the importance of the move saying it comes against a backdrop of the proliferation of “deep synthesis services…used by some criminals to produce, copy, publish and disseminate illegal information; slander and degrade people’s reputation, honour; as well as impersonating others’ identities to commit fraud and other illegal acts.”
Meanwhile, another aspect of the proposed rules highlighted by state-owned media outlet ECNS states that access to deepfakes or other AI-generated synthetic media content will be subject to identity verification processes.
This point requires deepfake content providers to carry out ID checks for users before allowing them access to their technologies or deepfake products.