New transparency obligations under the European Union's AI Act take effect on Sunday, requiring organisations and professionals to clearly label AI-generated and deepfake content distributed as part of their professional activity. Companies must also ensure chatbots disclose to users that they are interacting with AI. AI-generated texts on matters of public interest must be flagged unless editorially reviewed.
The rules, stemming from legislation that entered into force in August 2024, exempt artistic, satirical, fictional, and creative works, as well as texts only lightly edited with AI assistance. Violators face financial penalties, though existing AI systems have until December to adapt, with machine-readable watermarking requirements also deferred to that date.
Major platforms have moved ahead of the deadline: TikTok says over three billion pieces of content already carry AI labels, Meta has deployed an 'AI Info' tag on Instagram and Facebook, and Google has signed the EU's AI transparency code of conduct. The broader legislation classifies AI systems into four risk tiers and bans applications including social scoring and workplace emotion recognition.