The act bans the use of AI programmes that exploit human vulnerabilities, such as the use of subliminal techniques and social scoring for public and private purposes, as used in China to reward or punish individuals for their behaviour.