EU delays AI compliance
The European Parliament endorsed a negotiating position on an 'AI Omnibus' package but agreed to push the application of key high‑risk system requirements and compliance deadlines into 2027. At the same time lawmakers fast‑tracked bans on clearly harmful apps — like 'nudifier' tools — while industry groups warn the broader compliance burden could top €500 billion a year. (pinsentmasons.com) (europarl.europa.eu) (dqindia.com)
European Parliament plenary adopted its negotiating position on the Digital Omnibus on 26 March 2026 by 569 votes in favour, 45 against and 23 abstentions. (prod.iapp.org) The adopted text fixes application dates: core obligations for high‑risk AI systems would apply on 2 December 2027, sectoral‑law‑covered AI systems on 2 August 2028, and watermarking/transparency rules for AI‑generated content on 2 November 2026. (freevacy.com) The plenary position follows a joint report from the Internal Market (IMCO) and Civil Liberties (LIBE) committees that the committees adopted on 18 March with a committee vote of 101 in favour, 9 against and 8 abstentions. (ppc.land) The Council of the EU had already adopted its own negotiating stance on 13 March 2026, leaving trilogue talks between Parliament, Council and Commission as the next formal phase. (consilium.europa.eu) A coalition led by DIGITALEUROPE and 44 other European and national industry associations published a joint statement backing the Omnibus’s simplification aims, while a circulated industry letter cites broader EU regulatory compliance costs at around €500 billion per year. (digitaleurope.org) Officials and legal advisers say trilogue negotiations are expected to begin imminently and that some streamlined measures could enter into effect as early as 2 August 2026, while specific high‑risk obligations would await the December 2027 timeline set by Parliament. (pinsentmasons.com)
