EU AI Act delays let high-risk systems evade regulation

Delays to key provisions of the EU AI Act, combined with legal non-retroactivity, may leave some high-risk AI systems unregulated for a long time. The AI Omnibus position adopted by the European Parliament and Council postpones the core obligations for high-risk systems to December 2027 or even August 2028; under Article 111, systems placed on the market before the new deadlines do not need to comply unless substantially modified, while systems used by public institutions are deferred to December 31, 2030. Former AI Act co-negotiator Laura Caroli cited recruitment AI as an example, saying that if it is placed on the market before December 2, 2027, it may remain permanently outside the Act. Bram Vranken of Corporate Europe Observatory believes this will push companies to rush high-risk systems to market before the end of 2027. The Omnibus also proposes banning nudifier apps, but only for non-consensual content, and consent verification itself may pose privacy risks.