EU bans sexualized AI deepfakes in sweeping new deal lawmakers say goes too far

High risk AI rules now start later than expected

EU lawmakers and member states have agreed to ban AI systems that generate sexualized deepfakes, targeting humiliation and exploitation. The rollout of high-risk AI rules has been pushed to late 2027 and 2028, aiming to give businesses time and encourage innovation. The bloc is also tackling cybersecurity risks tied to increasingly advanced AI models.