European Union: European Commission adopted guidelines on scope of obligations for general-purpose AI models established by AI Act

On 18 July 2025, the European Commission published the Guidelines on the scope of the obligations for providers of general-purpose AI models under Regulation (EU) 2024/1689 (AI Act), ahead of the entry into application of Chapter V of the AI Act on 2 August 2025 (C(2025) 5045 final). These guidelines clarify the obligations applicable to providers of general-purpose AI (GPAI) models, including those with systemic risk, and aim to assist actors in determining whether a model qualifies as a GPAI model, whether they qualify as a provider placing such a model on the Union market, and whether exemptions apply under the AI Act. The guidelines define a general-purpose AI model as one displaying significant generality, capable of competently performing a wide range of distinct tasks, typically trained using over 10²³ floating point operations (FLOP) and capable of generating language, text-to-image, or text-to-video outputs. Providers of such models must ensure compliance with transparency, documentation, and copyright requirements, and, where systemic risk exists, determined by thresholds such as 10²⁵ FLOP or designation by the Commission, additional obligations apply, including model evaluation, incident reporting, and cybersecurity risk mitigation. Exemptions are outlined for open-source models provided under non-monetised, fully open conditions, unless the model poses systemic risks. The Commission, through the European Artificial Intelligence Office, holds exclusive enforcement competence and may use adherence to codes of practice as a means of demonstrating compliance. These guidelines are non-binding but set out the Commission’s interpretative framework for enforcement and will be subject to review.