The move underscores Beijing’s attempts to regulate the rapid development of user-centric artificial intelligence, boosting security and tightening ethical conditions.
The proposed regulations would apply to AI products and services that represent simulated human personality traits, thinking patterns and communication patterns and engage in emotional interaction with users through text, images, audio and video, or otherwise.
The draft envisages an approach that will require service providers to warn users of excessive use and intervene when they show signs of addiction.
According to the draft, service providers would have to take security measures as long as the product is in use and establish systems for algorithm auditing, data security and personal data protection.
The draft also foresees possible psychological risks. Service providers will be expected to recognize the user’s condition and assess their emotions and level of dependence on the service. If a user is found to be exhibiting extreme emotions or addictive behavior, the service provider must take certain measures and intervene, the draft says.
The measures set limits on content and conduct, stipulating that services may not generate content that threatens national security, spreads rumours, or promotes violence and obscenity.
(Vijesti.ba)
