The European Commission will start enforcing parts of the EU’s Artificial Intelligence Act from 2 August 2026, including new transparency rules requiring some AI systems to tell users when they are interacting with AI and when content has been generated or altered.
Chatbots and other interactive AI tools covered by the rules will have to make clear that users are dealing with AI rather than a human, the Commission said in a statement on Friday.
Deepfakes — images, videos or audio edited or generated using AI — will have to be labelled, while AI-generated or altered content will also need machine-readable markings designed to make it easier to detect.
The Commission also published a first list of more than 180 organisations that have signed a Code of Practice on transparency of AI-generated content.
Who enforces the rules, and what they cover
Enforcement will be led by the Commission’s AI Office working with national authorities, the Commission said.
The AI Office will be able to enforce the Act’s rules for providers of general-purpose AI models — systems that can be used for many different tasks and built into a wide range of tools and services.
The rules also cover the most advanced general-purpose AI models that may pose “systemic risks”, with extra obligations for providers to address risks of large-scale harm, including risks linked to chemical, biological, radiological and nuclear incidents, loss of control, cyber offence, harmful manipulation and threats to fundamental rights.
Providers of general-purpose AI models will be required to document certain information and provide it to competent authorities or downstream providers, and to put in place a copyright policy and publish a sufficiently detailed summary of the content used to train their models.
Responsibility for enforcing the transparency rules and banned AI practices will be split between bodies, with the AI Office covering AI systems offered by the same provider as the underlying general-purpose model and systems integrated into very large online platforms or very large online search engines designated under the Digital Services Act.
National competent authorities will enforce the rules for other AI systems, while the European Data Protection Supervisor will oversee AI systems used by EU institutions, bodies and agencies.
The Commission said the AI Office has appointed Professor Alessandro Abate of the University of Oxford’s Department of Computer Science as Lead Scientific Adviser, and that enforcement work will be supported by a Scientific Panel made up of 60 independent AI experts, which has held its first meeting.
The EU executive added that it has launched a complaints tool and a whistleblower tool for reporting alleged infringements of the AI Act, along with a channel for downstream providers building AI systems on general-purpose models to report suspected violations by the providers of those models.
Separately, the Commission said an “AI Omnibus” has postponed the application of rules on high-risk AI systems to 2 December 2027, and delayed rules for high-risk AI systems integrated into regulated products to 2 August 2028.
New prohibitions introduced by the Omnibus covering AI systems that generate non-consensual sexually explicit content and child sexual abuse material will apply from 2 December 2026.
