EU to apply AI Act transparency and GPAI rules from Aug 2
This digest was compiled by AI from multiple sources — links to the originals are below.
The European Commission and national authorities will begin enforcing AI Act transparency rules on August 2, 2026, requiring chatbots to disclose AI interaction and deepfakes to be labeled. The same day, the Commission gains oversight powers over general-purpose AI models, with extra obligations for the most powerful systems. A code of practice signed by over 180 organisations was published alongside the new rules.
Transparency Requirements
From August 2, chatbots and interactive AI systems must inform users they are dealing with an AI, not a human. AI-generated or modified images, video, and audio (deepfakes) must carry clear labels. All AI-created content must include machine-readable tags to identify its origin. The measures aim to reduce deception risks, increase transparency, and provide businesses with clear compliance guidelines.
Oversight of General-Purpose AI
The European Commission gains authority on August 2 to enforce AI Act rules for general‑purpose AI model suppliers. For models posing systemic risks, developers must assess and mitigate threats including cyberattacks, loss of control, dangerous manipulation, and chemical, biological, radiological, and nuclear risks. All suppliers must document model information, respect copyright, and publish detailed summaries of training data. Enforcement will be carried out by the EU AI Office, national competent authorities, and the European Data Protection Supervisor.
Enforcement Tools and Code of Practice
The Commission launched digital tools for citizens, companies, and whistleblowers to report suspected AI Act violations. At the same time, it published the first list of over 180 organisations that have signed the Code of Practice on AI‑generated content transparency. The EU had earlier opened a tender for up to seven AI gigafactories to boost competitiveness and technological independence.
What's Next
The AI Act obligations will be phased in gradually, with the next major deadlines in 2027 for high‑risk systems. It remains unclear how effectively the EU can enforce the rules across the diverse AI ecosystem, particularly for open‑source models.
2 sources
EU to apply AI Act transparency and GPAI rules from Aug 2
