appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Meredith Chen

China’s top court has issued new guidance on how judges should handle civil cases involving artificial intelligence, setting out rules on liability for deepfakes, voice cloning, hallucinations and violations of personal rights.
Although Beijing has yet to pass a dedicated AI law, the guidelines expand the country’s legal framework governing the technology.
The Supreme People’s Court’s 24-article opinion, released on Monday, instead provides guidance for handling AI-related disputes under existing laws, including the Civil Code, Cybersecurity Law, Copyright Law and Personal Information Protection Law.
The rules were the first of their kind and aimed to “draw clear legal red lines” while supporting innovation and development, the court said, aligning itself with the current national policy.
A major focus of the opinion was the use of AI to create deepfake images and videos, clone people’s voices and digitally “bring the dead back to life”.
The court said this could infringe legal protections covering aspects of a person’s identity and elements such as their image, voice and reputation.
Si Yanli, deputy director of the Supreme People’s Court’s research office, told a press conference that using AI without consent to generate and publish an identifiable digital likeness of someone could infringe their rights.


Using a person’s voice as training data without consent to generate a recognisable synthetic voice by imitating their timbre, tone and speaking style could also be an infringement.
The top court also cited media reports about “one-click undressing” tools used to create sexualised images of people, saying victims could apply for an injunction if failing to stop the spread of sexualised material could cause irreparable harm.
Courts could also order people to stop spreading sexualised material or require internet platforms and AI service providers to act to prevent further harm, the guidelines said.
The court said the damage from such abuse was difficult to reverse because AI-generated content could be produced at low cost and spread rapidly.
“The traditional approach of seeking redress through litigation after the fact is not conducive to providing timely and effective protection for victims’ legitimate rights, making the proper application of personal-rights injunctions particularly important,” the court said.
It also warned that the serious misuse of deepfake technology, including for fraud, defamation, personal-information violations and the production or distribution of obscene materials, could result in criminal liability.
The guidelines also set out when generative AI providers could be held responsible for content that damaged people’s reputations or breached their privacy. They would not be expected to screen every prompt and output in advance, but could be held liable if they failed to act after being notified of infringements.
Meanwhile, users who deliberately used AI to generate such content could also be held liable.
Other areas covered in the guidelines include consumer disputes, such as “big-data price discrimination”, where algorithms are used to offer customers different prices and terms for the same products, and AI-generated impersonations of celebrities in live-stream sales.
Consumers will be able to seek punitive damages when such impersonation constitutes fraud.
The guidelines were less clear on copyright disputes and did not rule on whether AI-generated content could itself qualify for copyright protection or whether using copyrighted works without permission to train AI models constituted infringement.


Instead, courts were instructed to consider factors including the type of AI service, the source of training data, the parties’ participation and measures taken to prevent infringement when determining liability in such cases.
Wang Liming, a vice-president of the China Law Society and a professor at Renmin University, wrote in an article published by the official People’s Court Daily on Tuesday that China should adopt a different approach to both the European Union and the United States.
He said the EU’s approach prioritised security and strong regulation, while the US placed greater emphasis on development and lighter regulation.
“China should avoid the constraints of both models and seek a balance between AI development and security,” he wrote.
He said the court’s guidelines reflected a distinctly Chinese approach to AI judicial governance, which made support for innovation and development a fundamental principle.
In contrast, the latest additions to the EU’s AI Act, which came into force last month, will oblige tech firms to make AI-generated content identifiable.
Meanwhile, those using deepfakes will also be required to disclose that the content is artificial.


Why Russia’s AI-powered economic future lies in its Far East

People’s Daily rejects US claims of malicious AI distillation, warns of countermeasures

First update made to ISO 9001 in a decade, with China shaping AI, cybersecurity standards

Unlikely allies Sanders and Bannon call for stronger AI oversight

AI: Policy

China’s highest court sets out new AI red lines with rules on deepfakes and privacy

The guidelines include measures designed to protect people’s privacy and image. Photo: Getty Images

Why China is being urged to study Ukraine’s wartime PR blitz in the AI age

Ukraine has used a series of strikes on Russian airbases as part of its information war against Moscow. Photo: EPA-EFE

China will never allow any country in this world to achieve AI dominance: Victor Gao

The United States is “weaponising and militarising AI”, according to Victor Gao. Photo: TNS

News

China

Politics

New guidelines provide a broader legal framework for everything from privacy breaches and sexual content to fraud and consumer protection


New guidelines provide a broader legal framework for everything from privacy breaches and sexual content to fraud and consumer protection.


United States

Artificial intelligence (AI)

China’s highest court sets out new AI red lines with rules on deepfakes and privacy

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}New guidelines provide a broader legal framework for everything from privacy breaches and sexual content to fraud and consumer protection

New guidelines provide a broader legal framework for everything from privacy breaches and sexual content to fraud and consumer protection