European Commission Confirms First AI Act Enforcement Requests for Information to More Than 30 AI Companies

On September 1, 2026, the European Commission confirmed that the European AI Office had initiated its first active enforcement inquiries under Regulation (EU) 2024/1689 (the AI Act), sending Requests for Information (RFIs) to more than 30 AI companies. One group of requests addresses safety and security matters, including general-purpose AI models and the most advanced models that may present systemic risks.

Another group concerns copyright and transparency obligations, including documentation and information on training content. The Commission did not publicly identify the recipients.

RFIs are an investigative compliance-verification measure and do not constitute a finding of infringement. The action does not amend the AI Act or create new legislative duties; it demonstrates that the AI Office has moved from implementation guidance to active documentary supervision.

Under the Commission's official enforcement framework, RFIs may be used to verify compliance, and failure to respond or the provision of incorrect, incomplete or misleading information may lead to fines, depending on the form and legal basis of the request.

Impact

Impact assessment: Medium.

Key reasons: this is the first confirmed use of the AI Office's RFI powers and signals active documentary review of safety, security, transparency and copyright compliance.

Who is affected?

Direct recipients, other GPAI providers, providers of models with systemic risk, downstream AI providers and their legal, safety, cybersecurity and documentation teams.

Conclusion:

the action creates no new legislative requirement, but covered providers should ensure that AI Act documentation is current, internally consistent and ready for production within an RFI deadline.