China Enacts Groundbreaking Regulations To Vet Generative AI — Will Other Nations Follow Suit?

Some of the world’s earliest and most detailed governmental regulations for generative artificial intelligence (AI) just went into effect in China on Aug. 15. These “Generative AI Measures” are intended to balance state control with state endorsement of the emerging technology in order to encourage its use while simultaneously protecting intellectual property (IP) rights and user privacy concerns. According to Article 1 of the regulations, China put these measures in place “[i]n order to promote the healthy development and standardized application of generative artificial intelligence, safeguard national security and social public interests, and protect the legitimate rights and interests of citizens, legal persons, and other organizations.”

The regulations apply to the use of generative AI technology to provide services for generating text, pictures, audio, video and other content. While China attaches “equal importance to development and security, promoting innovation and combining law-based governance,” it also imposed these regulations to ensure that generative AI services “abide by laws and administrative regulations” and “respect social morality and ethics.” With regard to IP rights, AI services must “[r]espect intellectual property rights” and “the legitimate rights and interests of others … and must not infringe on the portrait rights, reputation rights, honor rights, privacy rights, and personal information rights of others.”

Under the newly adopted regulations, all generative AI providers must register their services and submit these services to a security review by the Cyberspace Administration of China, the state cyberspace and information department, prior to their public release. As part of the security review, the data and foundation models employed by AI to generate content will be vetted to make sure they are based on legitimate sources and that they do not violate IP rights or leverage protected material without permission.

The regulations also require that all content created by generative AI be properly marked or labeled as such to prevent any generative AI material from being mistaken as human-authored content. Other articles in the regulations relate to what information the providers must supply to their users (e.g., guides and trainings), requirements for preserving and protecting user information, prohibitions against discriminatory uses of AI, provisions for fielding user complaints and making related updates to the system, and requirements for monitoring and controlling content generated by AI, including preventing and correcting any violations of privacy or likeness rights.

Penalties for violations of the regulations include warnings followed by suspension of the generative AI system (after a three-month warning or correction period); other penalties in accordance with applicable laws and regulations, such as the Chinese Cybersecurity Law, Data Security Law, Personal Information Protection Law and Science and Technology Progress Act; and criminal prosecution if crimes are committed.

It will be interesting to see how these new regulations impact generative AI and whether the United States and other nations will adopt similar generative AI measures.