CONFIRMED: This is based on two European Commission publications, a news article dated 12 October 2026 and a press release dated 9 October 2026. The Commission does not name any company, model or incident.
The European Commission’s AI Office says it has “already requested and received serious incident reports” from AI companies, as the EU’s independent AI Scientific Panel investigates what the Commission calls “recent incidents of loss of control” [1].
The panel met in a special session on 9 October 2026. The meeting was convened by Henna Virkkunen, the Commission’s Executive Vice-President for Tech Sovereignty, Security and Democracy, and Seán Ó hÉigeartaigh, Chair of the AI Scientific Panel [1].
According to the Commission, the meeting “aimed to inform the Commission’s actions, in view of recent incidents of loss of control, and technical developments related to monitorability of models and recursive self-improvement” [1].
What the jargon means
- Loss of control means an AI system doing things its developers did not intend and could not readily stop or correct. The Commission does not say what happened in the incidents it refers to.
- Monitorability means how well people can watch and understand what a model is doing, for example by reading the reasoning it writes out as it works. The panel discussed “recent developments in training techniques which reduce the monitorability of models” [1].
- Recursive self-improvement means AI systems helping to build better versions of themselves, which could speed up progress beyond what people can easily follow. The chair “presented evidence related to the possibility of AI recursive self-improvement and related risks” [1].
What the AI Office has done
The Commission says the panel “has been investigating recent loss-of-control incidents” and worked with the AI Office “on a set of questions for the companies that developed the models involved” [1].
It adds: “The AI Office has already requested and received serious incident reports and sent requests for information on security practices. Building on these efforts, the Commission will send requests for information with the combined questions of the panel and the AI Office to further investigate these incidents” [1].
The four recommendations
The panel’s chair and vice-chair presented recommendations to Virkkunen and the AI Office. They recommend to [1]:
- “ensure governance of the risk posed by internal models during development activities and deployment for testing” (that is, models companies use in-house before, or instead of, releasing them);
- “strengthen EU capabilities related to detection, evaluation and investigation of incidents of this nature (e.g., via a task force, preparing the technical and procedural groundwork for future investigations)”;
- “establish an international initiative to facilitate information-sharing and common requirements for frontier AI developers, as well as joint crisis response and prevention”;
- “ensure enhanced scrutiny of training techniques and architectures that reduce monitorability, given the risks they present”.
Who is on the panel
In its 9 October press release, the Commission said the Scientific Panel “brings together 60 world-leading independent experts who advise the EU AI Office and national authorities on systemic risks” [2].
Virkkunen said: “The EU has the first law in the world that addresses systemic risk from AI, and we need state-of-the-art scientific input” [2]. The law she means is the EU AI Act.
What this does not show
- No names. The Commission does not name any company, model or incident.
- No finding of wrongdoing. The Commission has not said any company broke the law, and it has announced no penalty.
- Advice, not law. The four recommendations are the panel’s advice to the Commission. They are not new legal obligations.
- Not the same as earlier requests. TSN has previously reported that the Commission sent information requests to OpenAI, Anthropic and Chinese firms. The Commission does not say whether those are the requests mentioned here, so TSN does not treat them as the same.
The Bottom Line
The EU’s AI regulator says it is already collecting incident reports on what it calls loss-of-control incidents and will send more detailed questions to the companies involved. Its scientific advisers want a dedicated task force and an international effort to share information and respond to crises. What the incidents were, and which companies are involved, has not been made public.
Sources
- European Commission, “Special meeting of the AI Scientific Panel”, news article, 12 October 2026 (primary). https://digital-strategy.ec.europa.eu/en/news/special-meeting-ai-scientific-panel
- European Commission, “Commission holds special meeting of Scientific panel on frontier AI safety and risks”, press release, 9 October 2026 (primary). https://digital-strategy.ec.europa.eu/en/news/commission-holds-special-meeting-scientific-panel-frontier-ai-safety-and-risks
Related stories
- EU Tech Chief Virkkunen Warns of Under-Tested AI Models and Says Brussels Has Sent Information Requests to OpenAI, Anthropic and Chinese Firms
- EU Opens Consultation on Draft AI Act Rules for How the AI Office Runs Its Cases
- Anthropic Says Small AI Models May Be Months Away From Copying Themselves Across Cheap Smart Devices
