EU Tech Chief: EU AI Act Sufficient to Address Uncontrolled AI Risks

EU tech chief Henna Virkkunen said the bloc's existing artificial intelligence regulations are sufficient to address risks posed by uncontrolled AI agents. In an interview with Reuters on Friday, she stated that the safety and security of models with powerful capabilities is a hot topic internationally, and Europe is well-prepared in this regard. The EU Artificial Intelligence Act, which regulates by risk level and is among the world's strictest AI rules, was adopted two years ago. Virkkunen dismissed critics who argue the EU's AI rules are outdated, saying the Act covers the entire lifecycle of these models and that regulators are providing guidance on model evaluation to companies based on recommendations from 60 AI experts. This comes after related incidents at OpenAI and Anthropic triggered global concerns that uncontrolled AI agents could slip beyond human control. The European Commission had sent requests at the end of August to more than 30 unnamed AI companies for detailed information on their safety and security measures, and Virkkunen said those responses are being assessed. Anthropic CEO Dario Amodei previously wrote in a blog post that AI is developing too quickly and, if unconstrained, could "surpass our ability to understand and control the system."