The European Central Bank has given Europe's largest lenders until October 31 to submit plans for countering cybersecurity risks posed by advanced artificial intelligence models. The ECB's Supervisory Board sent letters to 110 directly supervised banks outlining the requirement, with Chair Claudia Buch warning that cutting-edge AI represents a long-term shift in the threat landscape. The joint directive, issued with the European Systemic Risk Board, followed the ESRB's decision in June to elevate systemic cyber risk from "elevated" to "severe."
The warning coincides with growing scrutiny of frontier AI capabilities. Anthropic's Mythos model has demonstrated significant effectiveness in identifying and exploiting vulnerabilities in computer systems, though the company released a public version with built-in safety safeguards last month. The U.S. Commerce Department briefly restricted foreign access to Mythos 5 in June before reversing the order.