EU Tech Chief Says AI Act Can Handle Rogue AI Agents

European Union technology chief Henna Virkkunen says the bloc is well prepared to deal with rogue AI agents. Her comments follow incidents at OpenAI and Anthropic that have raised global fears of agents slipping beyond human control. She said the safety and security of highly capable models is a hot topic worldwide, and that Europe is well equipped to handle it. Virkkunen rejected claims that the rules are outdated, noting that the AI Act covers the whole life cycle of these models.

How the AI Act Regulates Advanced Models

Adopted two years ago, the AI Act regulates by risk level and is among the strictest AI laws anywhere. Regulators are guiding companies on how to evaluate their models, using recommendations from 60 AI experts.

EU Safety Checks on AI Companies

In late August, the European Commission asked more than 30 unnamed AI companies for details on their safety and security measures. Virkkunen said she is now reviewing the replies.

Anthropic CEO Dario Amodei wrote in a Sept. 13 blog post that AI is advancing too fast and may outrun our ability to understand and control it. He tied the pace to recursive self improvement, where AI increasingly builds the next generation of AI.