EU AI Scientific Panel holds special meeting on frontier AI loss-of-control incidents

Archive entry: this covers an event from October 9, 2026, and we published it on November 8, 2026. The details come from sources dated at the time and may have changed since.

On October 9, 2026, Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy, and Seán Ó hÉigeartaigh, chair of the AI Scientific Panel, convened a special meeting of the panel, as described on the Commission's Special meeting of the AI Scientific Panel page. It was prompted by “loss-of-control” incidents and by technical developments affecting model monitorability and “recursive self-improvement”.

What the meeting covered

• The panel and the AI Office jointly drafted questions for the developers of the models involved. Neither Commission page named the models, the companies or the incidents, or gave incident dates.
• The chair presented evidence on the possibility of AI improving itself recursively and on the risks involved.
• The panel discussed training methods that make models less monitorable, and what this implies for future incidents and risk mitigation.
• The AI Office had obtained serious incident reports after requesting them and had sent information requests on security practices. The Commission said it would send further requests combining the panel's questions with the AI Office's.

Four recommendations from the panel's chair and vice-chair

• Govern the risks posed by “internal models” during development and when deployed for testing.
• Build EU capacity to detect, assess and investigate such incidents, possibly through a task force that prepares methods and processes for future investigations.
• Set up a global effort that helps frontier AI developers share information, agree common standards and coordinate on crises and prevention.
• Increase scrutiny of training methods and model designs that make systems harder to monitor.

What it meant for builders

Builders who depended on frontier models had a reason to note that the EU AI Office was collecting incident reports and questioning model developers about loss-of-control incidents and security practices. The Commission described the Scientific Panel as 60 independent experts advising the AI Office and national authorities on systemic risk, model classification, evaluation methods and cross-border market surveillance. The Commission published a second page on the same meeting.

Sources

• Special meeting of the AI Scientific Panel (European Commission (DG CONNECT), October 12, 2026, official)
• Commission holds special meeting of Scientific panel on frontier AI safety and risks (European Commission (DG CONNECT), October 9, 2026, official)

Written with AI assistance, strictly from the sources listed above. Before it went live, automated checks verified its dates, sources and every language edition.