China's Regulations on Human-like AI Interaction Services Take Effect
China has implemented new regulations for AI services that mimic human interaction, establishing safety responsibilities and operational limits for providers.
China's cyberspace administration and four other ministries have jointly released new regulations for the management of AI services that simulate human interaction, effective July 15.
These regulations apply to services where artificial intelligence emulates human personality traits, thinking patterns, and communication styles to offer continuous emotional interaction through text, image, audio, or video. Such applications include emotional care, companionship, and support. Services not involving continuous emotional interaction, such as customer service or knowledge-based queries, are excluded.
The rules aim to balance AI development with safety, promoting innovation alongside legal governance. They encourage service providers to expand applications in areas like cultural dissemination, care for children and the elderly, and support for vulnerable groups, while emphasizing the importance of security awareness and legislation.
The regulations explicitly prohibit six categories of activities, including content that harms national security, user health, or rights. Particular attention is given to minors, for whom virtual relatives or partners are forbidden. Parental consent is required for services aimed at children under 14. Additionally, requirements for safety assessments and algorithm registration are introduced, and providers are encouraged to use AI sandbox platforms for technological testing and development.