Europe's answer to AI regulation complexity is to just delay most of it

The European Commission, the European Parliament, and the Council of the EU have agreed on simpler rules for artificial intelligence.

The deal is based on the "Digital Omnibus on AI," a legislative package that bundles several amendments to the existing AI Act. Rules for AI systems in areas like biometrics, critical infrastructure, education, and migration won't take effect until December 2027, while rules for AI in products like "lifts" or toys are pushed back to August 2028. AI systems that generate sexually explicit content without consent, including so-called "nudification" apps, are now explicitly banned.

Small and medium-sized enterprises with up to 750 employees and 150 million euros in revenue will get some relief, including reduced registration and documentation requirements along with better access to regulatory sandboxes - test environments where companies can try out AI under real-world conditions.

Labeling requirements still kick in August 2026

The Commission is calling the result "innovation-friendly," but IT lawyer and Heise legal counsel Joerg Heidrich isn't so sure. He says pushing back the high-risk deadlines to December 2027 or August 2028 was inevitable, since the original August 2026 timeline was practically impossible to hit.

Heidrich points out that the labeling obligations under Article 50 of the AI Regulation stay largely the same. Companies will still need to label deepfakes and certain AI-generated text starting August 2, 2026. That said, the text labeling rule apparently only applies to fully automated content that no human has reviewed or edited. So in practice, its real-world impact could end up being pretty limited.

The proposal was introduced in November 2025 as part of the EU's simplification agenda. Parliament and the Council still need to formally sign off on the agreement. The stated goal is to boost Europe's competitiveness while still protecting its citizens.