European AI Office Requests Information from Over 30 Model Providers

The European AI Office sent its first requests for information to more than 30 providers of general-purpose AI models under Article 91 of the European AI Act, according to a research note by the Cybersecurity Alliance. The requests address how models are secured, independently evaluated, and monitored post-market release.

This step is the first of its kind in the enforcement roadmap after the European AI Act entered into force gradually. The requests target providers of general models that can be used in multiple applications, regarding which covered entities are required to disclose detailed information.

The note clarified that incomplete or unsatisfactory responses expose the company to fines of up to 15 million euros or 3% of its global turnover, whichever is higher. These figures make responding to requests a matter of financial risk rather than merely formal compliance.

The European Commission has not disclosed the names of the companies that received the requests, keeping their actual scope unknown to the public. The matter concerns any provider selling models or services based on them within the European Union, putting both regional and global companies under similar obligations when dealing with the European market.

What do these terms mean?

European AI Act: A set of European rules that classifies AI systems according to their risk level and imposes obligations on them. Article 91: A clause granting the European Office the authority to request information from model providers to verify compliance. General-purpose models: Powerful models that can be used in many applications and are not tailored for a single task, such as models that write text or code. Global turnover: A company's total revenue across all markets, used as a basis to calculate fines so that penalties remain effective for large corporations.