EU delays AI rules
The EU has postponed key parts of the AI Act, creating a regulatory gap that lets high‑risk systems dodge oversight — regulators also moved to ban fully AI‑generated visuals in official communications amid trust concerns. 78% of European firms say they're unprepared for compliance, and security experts warn the delay leaves Europe exposed while North America and Asia sprint on next‑gen AI hardware. (techpolicy.press) (aa.com.tr) (natlawreview.com) (digitimes.com)
European Parliament’s plenary on 26 March 2026 adopted its negotiating position on the Digital Omnibus on AI with 569 votes in favour, 45 against and 23 abstentions, formally sending the file into trilogue negotiations. (nicfab.eu) The Council of the European Union had set its negotiating position on 13 March 2026, leaving three‑way talks between Parliament, Council and Commission to produce a final text that still requires Council sign‑off before any deadline changes take legal effect. (dataguidance.com) Draft application dates in the Omnibus proposed by negotiators include a fixed 2 December 2027 start for high‑risk systems listed in the regulation and 2 August 2028 for AI embedded in regulated products, shifting the timetable that firms had been preparing for. (onetrust.com) Parliamentary amendments explicitly seek extended timelines for categories such as biometric identification and law‑enforcement AI, and the plenary text also supports new prohibitions including 'nudifier' apps and non‑consensual sexual content generation. (biometricupdate.com) All three main EU institutions—the European Commission, European Parliament and Council—have adopted internal rules banning fully AI‑generated images and videos in official communications while allowing AI for optimization (for example, image enhancement), citing authenticity and trust as the rationale. (politico.eu) Vision Compliance’s 2026 readiness analysis found 78% of assessed enterprises had “not taken meaningful steps” toward AI Act obligations and flagged three common gaps: 83% lacked a formal AI inventory, 74% had no designated compliance owner, and 61% had no process for the technical documentation required for high‑risk systems. (globalfintechseries.com) Industry analysts warn the regulatory delay widens a practical window of vulnerability as North American and Asian players accelerate next‑generation AI hardware deployment; CONTEXT/ Digitimes reports characterise Europe’s stance as a two‑year lag that could become a security liability according to senior analysts. (digitimes.com)
