EU Reaches Deal to Delay AI Act Deadlines and Simplify Rules

European Union negotiators agree on the first amendments to the AI Act, pushing back compliance deadlines for high-risk systems and AI transparency requirements. The Digital Omnibus package also includes targeted simplification measures to ease the regulatory burden on developers.

EU negotiators reach a provisional agreement on the Digital Omnibus package, introducing the first major amendments to the AI Act since its adoption in 2024. The updated framework focuses on pragmatic timeline extensions and targeted simplification measures to help companies manage regulatory burdens. This deal reflects a realistic approach to the challenges of implementing complex artificial intelligence regulations across the bloc.

The most significant change involves staggered delays for high-risk AI system obligations. Use-based systems see their deadlines pushed back by 16 months to December 2027, while product-regulated systems receive a one-year extension to August 2028. Additionally, transparency requirements for synthetic content already on the market face a four-month deferral, and member states get an extra year to establish national AI regulatory sandboxes.

These timeline extensions give standard-setting bodies like CEN-CENELEC crucial additional time to develop the technical standards and guidelines required for compliance. Industry stakeholders welcome the relief, as high-risk systems demand extensive testing, documentation, and third-party assessments. The temporary delays aim to ensure a smoother, more practical rollout of the landmark AI legislation without compromising its core safety objectives.

Read More at the original source →