EU tech chief says bloc 'well equipped' to fend off rogue AI risk

Virkkunen told Reuters that Europe is "well equipped" to manage the safety and security of highly capable AI models. She highlighted that the EU's AI Act, which categorizes AI based on risk levels, encompasses the entire lifecycle of these models. The regulations also involve guidance for companies on model evaluation and the incorporation of recommendations from a scientific panel of 60 AI experts.

She dismissed criticisms that the EU's AI rules are already outdated, stating that legislators have effectively anticipated future developments by mandating risk assessments, the use of external experts, and continuous model monitoring. In late August, the European Commission initiated information requests to over 30 AI companies, seeking details on their safety and security protocols, as well as transparency and copyright obligations. These inquiries could potentially lead to investigations and fines of up to 7% of a company's global annual turnover. Virkkunen confirmed that similar information requests were also sent to Chinese AI startups, noting that the most advanced models currently available come from both the United States and China.

Regarding a potential digital levy, Virkkunen indicated that the decision rests with EU governments, who are currently discussing new financing sources for the EU budget from 2028 to 2034.