AI Omnibus
EU legislative package amending the AI Act; splits compliance into two deadline tracks and grandfathers in-market AI systems for Article 50.
Last refreshed: 17 June 2026 · Appears in 2 active topics
What does the AI Omnibus grandfathering mean for broadcasters already using AI tools?
Timeline for AI Omnibus
Mentioned in: EU finalises its AI content-marking Code
Media's AI PivotMentioned in: France chairs G7 Digital Ministerial on 29 May
European Tech SovereigntyMentioned in: Mistral buys into the industrial stack
European Tech SovereigntyDelayed Annex III compliance to 2 December 2027; GPAI deadline unchanged
European Tech Sovereignty: AI Omnibus deal splits enforcement into two speedsReached provisional agreement in May 2026 grandfathering in-market AI systems to 2 December 2026
Media's AI Pivot: Brussels moves Article 50 to DecemberWhat changed in the EU AI Act under the AI Omnibus deal?
When does the EU AI Act apply to high-risk AI systems after the Omnibus deal?
Why did the EU delay AI Act enforcement for high-risk systems?
Background
The Digital Omnibus on AI (AI Omnibus) is an EU legislative package that amends the 2024 EU AI Act as part of the Commission's Omnibus VII simplification agenda. On 7 May 2026, the Council of the EU and the European Parliament struck a provisional agreement splitting AI compliance into two speeds: Annex III high-risk AI obligations moved from 2 August 2026 to 2 December 2027 for stand-alone systems and to 2 August 2028 for high-risk AI embedded in products, while the General Purpose AI enforcement Deadline was left unchanged at 2 August 2026.
Beyond the enforcement split, the Omnibus added a new prohibition against AI-generated non-consensual intimate content and child sexual abuse material, and extended lighter documentation and penalty treatment to small and medium enterprises. Formal Council and Parliament adoption is required before 2 August 2026 to prevent the original high-risk Deadline from triggering. In the context of media and synthetic content, the Omnibus also grandfathered AI systems already on the market before 2 August 2026, giving them until 2 December 2026 to meet the Article 50 machine-readable marking requirement. New AI deployments after 2 August face the requirement at launch with no grace period.
The AI Omnibus is a bellwether for how the EU balances innovation policy against its regulatory apparatus. The two-speed approach was a concession to industry pressure without abandoning GPAI enforcement, which matters most for frontier model providers (including Mistral AI, OpenAI, Google DeepMind, and Anthropic) operating in EU markets. As of 10 June 2026, no EU broadcaster has publicly signed the Article 50 Code of Practice the Omnibus's grandfathering window was designed to give them time to adopt.