EU AI Act Faces Delays and Pushback Midway Through 2025 Rollout
The first half of 2025 reveals significant implementation hurdles for the EU AI Act, including delayed codes of practice for general-purpose AI and growing calls to pause enforcement.
The EU AI Act is currently navigating a complex transition from legislative text to active enforcement as 2025 reaches its midpoint. As the world's first comprehensive legal framework for artificial intelligence, the law categorizes AI systems by risk level and imposes strict obligations on high-risk and general-purpose AI (GPAI) models. A structured timeline dictates that prohibitions on certain AI systems take effect in February 2025, followed by GPAI obligations in August 2025, and full application of remaining provisions by August 2026.
Despite this clear schedule, the European Commission indicates that several key deliverables face unexpected postponements. Most notably, the highly anticipated Code of Practice for GPAI models experiences significant delays as stakeholders struggle to reach a consensus. These setbacks create substantial questions regarding enforcement readiness and leave businesses seeking urgent legal clarity about their upcoming compliance requirements.
Adding to the uncertainty, a public consultation on high-risk AI systems is currently underway while increasing voices across the industry call for a formal pause in the AI Act's implementation. These tensions highlight the inherent difficulty of applying rigid regulatory frameworks to rapidly evolving technology. As the August 2025 deadline for GPAI provider obligations quickly approaches, organizations closely watch Brussels to see how regulators handle these mounting logistical and political pressures.