This feedback responds to the Commission’s draft implementing regulation establishing detailed arrangements for evaluations of general-purpose AI (GPAI) models and enforcement proceedings under Articles 92 and 101 of the AI Act. We welcome the Commission’s initiative in putting forward a procedural framework that provides much-needed legal certainty and helps operationalise the Act in several important respects. The provisions discussed below represent meaningful clarifications, supporting the EU’s capacity to scrutinise GPAI models and hold their providers accountable.

Our feedback is structured in two parts. We first identify the provisions we believe merit support, then present seven targeted recommendations for improvement. Throughout, our concern is to ensure that the procedural architecture established by this implementing regulation is robust enough to serve its purpose: enabling rigorous, independent evaluations of GPAI models and effective enforcement when providers fail to meet their obligations.