The EU AI Act was meant to reach its most demanding phase on August 2, 2026, when rules for “high-risk” AI systems kicked in. In May, Europe pushed that date back.

What changed

The European Commission proposed a “Digital Omnibus on AI” in November 2025 as part of a broader drive to simplify EU rules. On May 7, 2026 the European Parliament and Council reached political agreement, and the regulation entered into force on July 27. The main changes, according to the Commission:

  • High-risk AI in sensitive areas delayed. Stand-alone systems listed in Annex III, covering biometrics, critical infrastructure, education, employment, migration and similar uses, now have until December 2, 2027 instead of August 2, 2026.
  • AI in regulated products delayed further. AI embedded in products covered by existing EU safety law (Annex I), such as machinery or medical devices, moves to August 2, 2028.
  • A new ban. AI systems that generate non-consensual sexual or intimate content, or child sexual abuse material, are prohibited. Gibson Dunn reports a transition period to December 2, 2026.
  • Lighter rules for growing companies. Some privileges for small and medium-sized enterprises extend to small mid-caps, an EU-level regulatory sandbox is added and the AI literacy duty is simplified.
  • A stronger AI Office. It gains enforcement powers over AI systems built on general-purpose models and over AI in very large online platforms and search engines.

What did not change

This matters most for developers, because the delay is narrower than headlines suggested. According to Gibson Dunn:

  • General-purpose model obligations have applied since August 2, 2025 and are unchanged.
  • Transparency duties under Article 50, such as telling people they are talking to an AI and marking AI-generated content, apply from August 2, 2026, with a grace period to December 2, 2026 for machine-readable marking on systems already on the market.

Why it matters

  • More time, same obligations. Teams building hiring tools, credit scoring or biometric systems for the EU gain about 16 months. The requirements themselves, risk management, documentation, human oversight, are still coming.
  • Transparency is the near-term work. For most developers shipping chatbots or generative features in the EU, labelling and disclosure are what apply in 2026.
  • Europe’s simplification turn. The omnibus is the clearest sign yet that competitiveness concerns are reshaping EU tech regulation. It is unlikely to be the last adjustment.

Gibson Dunn’s analysis predates the final text; check exact grace periods against the Official Journal before relying on them.

Sources

  1. European Commission: press release IP/26/1024 on the AI Omnibus agreement (May 7, 2026)
  2. European Commission: EU agrees to simplify AI rules and ban nudification apps (May 7, 2026)
  3. European Commission: AI Omnibus enters into force (Jul 27, 2026)
  4. Gibson Dunn: EU AI Act omnibus agreement, postponed high-risk deadlines (May 27, 2026)

We link every claim to the reporting or official statement it comes from. Where a detail rests on one outlet’s reporting, we say so. Spotted an error? Reply to any newsletter and we’ll correct it.