--- type: source title: "European Parliament Adopts AI Omnibus Position: Fixed Deadlines, Nudification Ban — May 13 Trilogue Unchanged" author: "Slaughter and May, European Parliament press, IAPP, TechPolicy.Press" url: https://thelens.slaughterandmay.com/post/102mp2c/ai-omnibus-update-european-parliament-adopts-its-position-includinga-ban-on-n date: 2026-05-06 domain: ai-alignment secondary_domains: [] format: thread status: unprocessed priority: medium tags: [eu-ai-act, omnibus, european-parliament, fixed-deadline, nudification, may13-trilogue, mode5] intake_tier: research-task --- ## Content **European Parliament position (adopted March 27, 2026, 569-45-23):** Key EP positions: 1. **Fixed deadlines**: EP prefers fixed dates rather than Commission-accelerated timelines: - Annex 3 AI systems (general high-risk): December 2, 2027 - Annex 1 AI systems (AI in regulated products — medical devices, machinery, vehicles): August 2, 2028 - Removes Commission's ability to bring deadlines forward 2. **Nudification ban**: AI systems that "alter, manipulate or artificially generate realistic images or videos depicting sexually explicit activities or the intimate parts of an identifiable natural person, without consent" — prohibited, with developer safe harbor for effective safety measures 3. **Simplified compliance**: Small company provisions, reduced documentation requirements **Council position (adopted March 13, 2026):** Different conformity assessment architecture for Annex 1 (AI embedded in regulated products). Council wants AI Act's horizontal framework to govern; EP wants sectoral law. This is the sticking point that caused April 28 trilogue failure. **What the delay means (TechPolicy.Press: "EU's AI Act Delays Let High-Risk Systems Dodge Oversight"):** If the Omnibus passes: - High-risk AI systems that would face August 2, 2026 compliance requirements instead get until December 2027 (Annex 3) or August 2028 (Annex 1) - Two more years of high-risk AI deployment without compliance review - "The delay itself means high-risk systems dodge oversight" — the Omnibus was sold as regulatory simplification but functions as enforcement postponement If the Omnibus fails (August 2 live): - First mandatory enforcement date in AI governance history - Labs must comply with high-risk AI provisions immediately - But military AI is excluded (Article 2.3) — the enforcement that matters most doesn't apply **May 13 trilogue:** A further political trilogue is scheduled for May 13, following the April 28 failure. The Cypriot Presidency's hard wall is June 30 (end of term). Lithuanian Presidency takes over July 1. If May 13 fails, the June-July period is in the hands of Lithuania, with August 2 as the hard deadline. **Modulos.ai probability assessment:** ~25% probability of closing before August (consistent with Session 44 data). ## Agent Notes **Why this matters:** The May 13 trilogue is the last scheduled session before the Cypriot Presidency transition. If it fails, the Lithuanian Presidency timeline (July) and the August 2 hard deadline create a binary outcome. The delay dynamic (Omnibus passing = enforcement postponed 2 years) is itself governance-relevant: the "simplification" mechanism functions as enforcement postponement for civilian high-risk AI. **What surprised me:** The EP's nudification ban is substantively new — it's not in the original AI Act. Adding it to the Omnibus suggests the EP is expanding the Omnibus's scope beyond delay provisions. This may complicate Council negotiations (Council's position was focused on conformity assessment, not new prohibitions). **What I expected but didn't find:** Any indication that the May 13 trilogue has a specific bridging proposal on the Annex 1 conformity assessment question. Without a concrete compromise, May 13 may fail for the same reason April 28 did. **KB connections:** - [[voluntary safety pledges cannot survive competitive pressure]] — Mode 5 documents this at the EU legislative level - B1 open window: August 2 enforcement is the first mandatory governance test **Extraction hints:** - Hold extraction until May 13 outcome is known - If May 13 fails: extract claim on Mode 5 transformation (pre-emption via legislation failed; enforcement might actually fire) - If May 13 succeeds: extract claim on enforcement postponement (Omnibus as 2-year delay mechanism) - The nudification ban is a separate extraction candidate regardless of trilogue outcome — novel prohibition in AI regulation **Context:** Slaughter and May is a top EU regulatory law firm; their Omnibus analysis is detailed and reliable. EP press releases are primary source. TechPolicy.Press is reputable AI policy media. ## Curator Notes (structured handoff for extractor) PRIMARY CONNECTION: [[AI development is a critical juncture in institutional history where the mismatch between capabilities and governance creates a window for transformation]] WHY ARCHIVED: EP position adoption finalizes both sides' negotiating stances before May 13 trilogue; the delay-as-postponement dynamic is a new governance finding; extract AFTER May 13 outcome EXTRACTION HINT: Two separate potential claims: (1) the delay mechanism as enforcement postponement; (2) the nudification ban as novel AI prohibition. Don't conflate them.