China releases draft regulations to control AI that interacts like a human
China’s cyber regulator on Saturday released draft regulations for public comment that would strengthen oversight of AI services intended to mimic human personalities and elicit emotional responses from users.
The action highlights Beijing’s efforts to influence the quick adoption of consumer-facing AI by tightening ethical and safety standards.
The proposed regulations would cover AI services and products that are made available to the general public in China and that use text, images, audio, video, or other media to emotionally engage consumers while simulating human personality traits, thought processes, and communication styles.
The proposal outlines a regulatory strategy that would mandate that providers caution users against excessive use and step in when users exhibit symptoms of addiction.
According to the idea, service providers would have to set up procedures for algorithm evaluation, data security, and personal information protection, as well as take on safety duties throughout the product lifecycle.
The proposal also focuses on possible psychological hazards. It would be expected of providers to recognize user states, evaluate users’ emotions, and gauge how dependent they are on the service. It stated that providers should take the appropriate action to intervene if consumers are discovered to display intense emotions or addicted behavior.
The regulations establish red lines for conduct and content, mandating that services refrain from producing anything that promotes violence or obscenity, spreads rumors, or jeopardizes national security.
