EU to require labels for AI-generated content from Sunday

Companies in the European Union must label deepfakes and other AI-generated content from Sunday, August 2, as transparency rules under the bloc’s artificial intelligence law begin to apply.

The rules require AI systems such as chatbots to make clear to users that they are AI. Images, text and other content created with AI must also carry a label, including through watermarks or other markers that make synthetic content easier to detect.

The EU says the rules are meant to help people understand when online material is real and when it has been generated or manipulated by AI. An EU official said generative AI can produce disinformation at large scale and with high speed.

The requirements apply to content made for professional reasons. The EU says individuals using AI only in a personal capacity will not be affected. Text intended to inform the public on general interest issues will also need a label if it is created using AI without human editorial oversight.

Existing AI systems have until December 2 to adapt. The rules include exemptions for artistic, creative, satirical and fictional work.

Large technology companies have already introduced some AI labels. TikTok says more than three billion items on its platform already carry labels through creator tools and detection technology. Meta uses an “AI Info” label on Instagram and Facebook. Google has signed the EU code of conduct on AI transparency and says it is working with companies including Nvidia, OpenAI and Apple on digital tagging tools.

Google’s Karen Massin warned that overlapping AI labels and legal disclosures could make it harder for users to get clear context. Ashley Casovan of the International Association of Privacy Professionals told AFP that implementation would be difficult but said companies often work through compliance requirements.