EU Parliament co-rapporteurs are working toward a three-layer regulatory framework that applies specific obligations to general-purpose AI models, foundational models, and generative AI tools directly within the AI Act text, rather than deferring to future implementing rules as Member States prefer. The framework targets transparency and training-data documentation across the value chain, with a centralized enforcement office to minimize forum-shopping risk. Expected parliamentary votes in early June 2023 would set up trilogue negotiations with Council, with target completion by end of 2023 and enforcement likely not before 2025. This represents a material hardening of regulatory surface area for generative AI vendors and a structural preference to address liability and provenance throughout the supply chain rather than downstream only.
314 other tracked developments touching the same claims.