Mythos is a new AI model from Anthropic that stands out for its ability to program at a very high level and detect cybersecurity vulnerabilities and propose how to exploit them.

Reuters says that this combination has led supervisors and regulators to treat it as a potential new source of risk, especially for banking and infrastructure with legacy systems.

Supervisors at the European Central Bank (ECB) are about to question bankers about the risks that Anthropic's new artificial intelligence model could amplify cyberattacks, a source familiar with the situation told Reuters on Wednesday.

ECB supervisors are gathering information about the model with the intention of asking the banks under their supervision about their preparedness for this potential new source of risk.

Mythos's high-level programming capabilities have given it a potentially unprecedented ability to "identify cybersecurity vulnerabilities and devise ways to exploit them." 

US Treasury Secretary Scott Bissent and Federal Reserve Chairman Jerome Powell convened an emergency meeting last week with bank chiefs to warn them about the risks.

Alberto Musalem told Reuters that the development underscored the need for the US central bank to "rethink how we think about cybersecurity" and to consult with banks about their own "resilience and robustness against cyber risk in this new world."

British Technology Secretary Liz Kendall and Security Minister Dan Jarvis issued a similar warning to companies on Wednesday, stating that Mythos was "substantially more capable at detecting cybercrime" than any model previously tested by the government's AI Security Institute.

Bank of England Governor Andrew Bailey said this week that central banks and financial regulators must quickly grasp the implications of the new model.