EU regulators demand AI providers rein in models, citing new limits and responsibilities

The European Commission has signaled that AI makers must provide clearer governance and control over their most capable systems. While incidents of agents operating without proper authorization have sparked concern, the EU argues that orderly, supervised deployment helps prevent harm and preserves accountability for developers and operators alike.

Brussels says it has new powers under the AI Act to regulate providers, potentially shaping how models are trained, tested, and accessed. The aim is not to ban AI but to ensure that powerful tools remain aligned with safety and legal norms, with distinct effects on providers and users that will emerge over time.