China's Supreme People's Court released its first judicial guidelines for handling artificial intelligence disputes on September 7, addressing a range of emerging issues, including AI deepfakes and algorithmic price discrimination.
On deepfakes and digital resurrection
On deepfakes, the guidelines state that using a person's name or likeness without consent to generate a recognizable virtual avatar amounts to an infringement of that person's rights to their name and portrait.
The same applies to voice cloning: training a model on someone's voice without permission to produce a recognizable imitation is also a violation of personal rights, according to the guidelines. Anyone who uses a cloned avatar or synthetic voice to spread false statements or otherwise damage a person's reputation can be held liable.
The guidelines also address a more unusual trend: the use of AI to "revive" the dead. Where someone creates or uses a digital likeness of a deceased person without authorization, and it damages that person's name, image or reputation, surviving relatives can seek damages under China's Civil Code, the court said, adding that such claims will be supported.
On doxxing and privacy online
Meanwhile, the guidelines take aim at AI-assisted doxxing and so-called "human flesh searches," a form of crowdsourced online exposure of personal information on social media. Using AI to track and analyze a person's public data, such as phone numbers and social media accounts, to unearth private information, then leaking or exploiting it, will be treated as a violation of privacy rights. The same goes for using AI tools to secretly film, monitor or eavesdrop on someone's private life.
On algorithmic price discrimination and fake endorsements
The court also moved to curb "big data profiling," a widely criticized practice in which companies use algorithms to charge different prices to different customers for the same product or service. Businesses that harm consumers through such differential pricing can now be held liable for damages, the guidelines state.
AI-generated celebrity endorsements used to defraud consumers were also singled out. Where such endorsements are found to be fraudulent, consumers pursuing punitive damages will have the backing of the courts, according to the document.
On self-driving liability
The guidelines also address liability in accidents involving self-driving and driver-assist vehicles, an increasingly contentious issue as the technology spreads across China's auto market. Where a vehicle defect and driver error combine to cause an accident, victims can pursue claims against both the driver and the vehicle's manufacturer or seller simultaneously, the guidelines state.
Manufacturers and dealers that make false or misleading claims about a vehicle's automation level or performance can also be held civilly liable if consumers are harmed as a result.
The Supreme People's Court acknowledged the guidelines could not cover every scenario, saying it would continue to refine its approach through future case rulings, judicial oversight and cooperation with other agencies at home and abroad.
Reporter | Chen Siyuan
Photo | VCG
