EU adopts new rules to manage risks from rogue AI agents

The European Union has adopted new regulations establishing mandatory risk management protocols for autonomous AI agents operating without direct human supervision. The legislation requires developers to implement real-time monitoring, automatic shutdown capabilities, and liability insurance for systems capable of independent decision-making in critical sectors. EU officials stated the rules address emerging safety gaps in agentic AI that existing frameworks do not adequately cover. Technology companies must comply within eighteen months or face penalties up to seven percent of global turnover. Industry groups have expressed concern that compliance costs may disadvantage smaller European startups against well-capitalized international competitors. Consumer advocacy organizations praised the measure as necessary protection against uncontrolled automation. The regulation distinguishes between supervised tools and autonomous agents, applying stricter requirements only to systems demonstrating independent goal-pursuit behavior. Legal experts anticipate this framework will influence pending AI legislation in Asia and Latin America.