The AI Office has sent its first requests for information to more than thirty companies

Less than a month after the Commission's set of powers over providers of general-purpose AI models became applicable, the AI Office sent its first requests for information under Article 91 of the AI Act to more than thirty companies active in AI. Commission spokesperson Thomas Regnier confirmed this at the midday briefing on 1 September 2026, noting that it followed an announcement by Executive Vice-President Henna Virkkunen and that these are the very first enforcement steps on the Commission's side. The requests cover two areas: the first concerns safety and security, including for general-purpose and the most advanced AI models, the second copyright and transparency and went to a different set of companies. The Commission is not naming the recipients – according to the spokesperson precisely because these are simple requests for information and a first enforcement step, while the dialogue with all companies continues.

This is a preparatory step that does not in itself open proceedings; under Article 91(4) a request must state its legal basis, its purpose, the information required, the time limit and the fines provided for in Article 101, which for non-compliance or for incorrect, incomplete or misleading information may reach 3 % of total worldwide annual turnover or EUR 15 million, whichever is higher.

Asked about recent cybersecurity incidents, the spokesperson declined to confirm who had received a request, but said separately that the Commission had had recent exchanges with Anthropic and OpenAI, including on those risks. It does not follow from the briefing that those two companies are among the recipients.