China's regulation on AI companions takes force
15 July marked the effective date for China's new requirements and potential liability on providers of AI-powered "emotional interaction services," marking the latest jurisdiction to tackle companion and chatbot harms.
Contributors:
Joe Duball
News Editor
IAPP
Rules around artificial intelligence companion and chatbot rollouts are making waves around the world, with China the latest to clamp down on potential harms stemming from human interactions with generative AI models. China's Interim Measures for the Administration of Anthropomorphic AI Interaction Services took effect 15 July, establishing standards and limits around model behaviors while holding providers liable for noncompliance.
The measures mark one of the first instances of a comprehensive national policy to rein in potential companion and chatbot issues. The EU AI Act will require companion and chatbot providers to begin disclosing their nonhuman nature 2 Aug., while a range of U.S. state legislatures have passed similar laws on disclosure and user safety.
China's framework defines companions — or "emotional interaction services" — as technologies that "simulate the personality traits, thinking patterns, and communication styles of natural persons" while providing "emotional care, companionship, support, and other interactive services provided through text, images, audio, video, and other forms." There are a range of exempted technologies, including services used as work assistants or those used in education and scientific research contexts.
Provider requirements are highlighted by mandatory emotional distress detection, crisis intervention and curbs on addictive use. The measures also include personal data rights, including prohibitions on using sensitive personal data from companion interactions to train AI models.
The Cyberspace Administration of China will oversee compliance and enforcement. Penalties for noncompliance and refusal to correct cited violations include fines up to CNY100,000 for simple violations while a maximum CNY200,000 penalty applies to instances of "harm to citizens' life and health and safety is accompanied by harmful consequences."
Rimon Law Partner Sarah Zhao recently analyzed the measures alongside other recent regulatory rollouts comprising China's expanded AI oversight strategy.
