The proposed regulations target artificial intelligence products and services that replicate human-like thinking patterns and communication styles, engaging users on an emotional level. Such interactions may involve text, images, audio, video, and other forms of content.
The document is designed to tighten safety and ethical requirements for providers of these services.
Under the draft rules, service providers would be required to:
-
take responsibility for security throughout the entire lifecycle of their products;
-
implement algorithm auditing systems to protect data and personal privacy;
-
assess users’ emotional state and level of psychological dependence on the service, and intervene when necessary
The proposed measures also prohibit AI services from generating content that threatens China’s national security, spreads rumors, or promotes violence and obscenity.
China’s Ministry of Science and Technology introduced a set of ethical principles for AI regulation as early as 2021, emphasizing user rights protection and risk prevention. Key principles included ensuring AI systems are controllable and reliable, fair and impartial, and improving overall ethical awareness.
Two years later, the Cyberspace Administration of China issued rules governing generative AI algorithms. Alongside requirements such as content accuracy and respect for intellectual property, the regulations explicitly banned content that incites actions undermining the state system.
