EU says its AI Act covers emerging rogue-agent risks

European Commission executive vice-president Henna Virkkunen said the EU's AI Act provides a framework for addressing risks from increasingly capable AI, including systems that behave unexpectedly or outside their operators' intentions. The remarks followed new concerns about the autonomy and safety of frontier models.

The Commission has sought information from AI developers about safeguards and compliance. The debate is moving beyond whether models can produce harmful text toward testing, monitoring and accountability for systems capable of taking actions.