The European Central Bank highlights risks from advanced AI tools like Anthropic’s Mythos, urging immediate action despite limited access.
The European Central Bank has called on euro area banks to accelerate preparations for cyberattacks powered by advanced AI models, including Anthropic’s Mythos. ECB supervision vice chair Frank Elderson emphasized that lack of direct access to the tool does not justify delays, warning that future AI models could enable even more aggressive threats.
Elderson noted that U.S. banks with early Mythos access are already addressing vulnerabilities flagged by the system. Cybersecurity experts view Mythos as a major challenge to banking infrastructure, prompting regulatory warnings. The ECB itself faces limitations due to restricted access to the model, despite ongoing studies into defensive measures.
The ECB’s warning follows reports that supervisors will assess banks’ readiness for AI-driven cyber risks. Regulators globally are increasingly focused on the potential for AI to disrupt financial stability, with policymakers urging proactive measures to mitigate emerging threats.