The European Commission has held a special meeting of its Scientific Panel on artificial intelligence to address recent loss-of-control incidents involving frontier AI models.
The panel, working together with the Commission's AI Office, has prepared a set of questions to put to the companies that developed the models implicated in these incidents. The meeting was referenced in remarks attributed to Executive Vice-President Henna Virkkunen.
Details of the specific incidents under review, the models involved, and the companies that will be approached have not been disclosed in the Commission's announcement.
The Scientific Panel on AI was established to provide independent technical advice to the Commission and the AI Office, including on systemic risks associated with general-purpose AI models.
For providers of general-purpose AI models with systemic risk, this development is a reminder that regulatory scrutiny under the EU AI Act extends beyond paper compliance to active monitoring of real-world incidents.
Organisations developing or deploying frontier models should, as a compliance matter, ensure that incident reporting and risk-tracking processes are robust enough to respond quickly to regulator enquiries.
Veritome will continue to monitor developments from the Scientific Panel and the AI Office as further information becomes available, including any formal requests for information issued to developers.