EU begins enforcing AI Act transparency and deepfake rules

The European Union began enforcing key provisions of its landmark AI Act on August 2, requiring companies to disclose when users are interacting with artificial intelligence and to label AI-generated or manipulated content, including deepfakes, with machine-readable markers.sifted+1

New Obligations Take Effect

Under the rules now in force, chatbots, AI agents, and digital avatars must clearly inform users they are dealing with an automated system rather than a person. Deepfake images, video, and audio must be visibly identified, and covered content must include technical markers that allow platforms and automated tools to detect it. AI-generated text published to inform the public on matters of public interest must also be disclosed when it has not received human review or editorial oversight.theaiinsider

General-purpose AI model providers face additional requirements: they must document their models and make that information available to regulators, establish copyright compliance policies, and publish sufficiently detailed summaries describing material used to train their systems. Developers of the most capable models must also address systemic risks involving cyberattacks, harmful manipulation, and loss of control.theaiinsider

The European Commission said more than 180 organizations have signed a voluntary Code of Practice designed to help companies apply the transparency rules.theaiinsider

Enforcement and Penalties

Responsibility for enforcing the rules is divided among the European Commission's AI Office, national regulators, and the European Data Protection Supervisor. Companies found in breach face fines of up to 3% of their annual global turnover.sifted+1

"The AI Act's transparency requirements are wider in scope than most organisations realise," said Marcus Evans, a partner at law firm Norton Rose Fulbright. "This isn't just an issue for the major AI developers — it can also affect organisations rolling out AI tools with their own branding."sifted

The rules are expected to increase pressure on US AI companies serving European customers. "Any lab with European customers now needs to be aware that if its model's outputs reach EU users, it's considered in scope," said Peter Van Dyck, partner at law firm A&O Shearman.sifted

What Comes Next

Enforcement is expected to build gradually, with regulators initially looking for companies to demonstrate active work toward compliance rather than immediately levying penalties. A 60-member Scientific Panel of independent AI experts will provide technical support, and the AI Office has appointed Alessandro Abate of the University of Oxford as its lead scientific adviser.theaiinsider+1

Other provisions remain on the horizon. The EU plans to introduce additional prohibitions on December 2, 2026, while requirements for high-risk AI systems have been delayed until 2027 and 2028.theaiinsider