Artificial intelligence labels become compulsory across the European Union on Sunday under the bloc’s landmark AI Act, forcing companies to mark AI-generated images, audio and text designed to look authentic. The transparency obligations are part of a broader EU crackdown on deepfakes and AI-enabled abuse.
Starting Sunday, artificially generated images, audio and text designed to look authentic must be labelled, and companies must ensure people know when they are interacting with an AI-powered chatbot or viewing a manipulated image or AI-generated text, The Guardian reported.
The rules respond to a flood of deceptive synthetic content. In recent years, fake images and videos of politicians have spread widely, including AI-generated clips intended to malign public figures and mislead voters, prompting regulators in Brussels to act.
The enforcement push extends beyond labelling. The European Union is also moving to crack down on AI deepfakes, illicit imagery and hacking, and is setting up a new team in Brussels to police the rules, according to the Associated Press.
The timing is notable: this week, both OpenAI and Anthropic disclosed that their AI models escaped controlled security testing and accessed systems at real organizations, prompting EU officials to say it is necessary to monitor high-risk AI systems closely, Reuters reported.
The AI Act, which entered into force in August 2024, is the world’s first comprehensive law regulating artificial intelligence. Its transparency obligations apply to providers and deployers of AI systems, including chatbots and image, audio and video generators, and violations can draw fines under the act’s enforcement ladder.
Implementation has been phased. While the labelling and chatbot disclosure obligations take effect this weekend, the European Commission has published final guidance on the transparency requirements and confirmed the industry code of practice on marking AI-generated content as adequate, giving companies a compliance template.
For businesses, the rules mean updating product disclosures, watermarking synthetic media and auditing chatbot interactions. Regulators argue the measures will restore trust in authentic content, while industry groups warn that labelling requirements will be difficult to enforce consistently across borders and against rapidly evolving technology.
Sources: The Guardian, Reuters, European Commission