Anthropic's latest AI model Claude Mythos has leaked despite being deemed too dangerous for public release. Financial institutions now face advanced AI-powered attacks capable of exploiting unknown vulnerabilities.
Anthropic describes Claude Mythos as highly risky. The model has already become available outside the company and can identify as well as exploit unknown security gaps when instructed.
Banks are especially vulnerable because they operate complex IT systems that combine old and new components. Professor Co-Pierre Georg from the Frankfurt School of Finance names three main threats: loss of sensitive customer data, disruption of critical systems, and subsequent loss of trust.
The German Finance Ministry is monitoring developments closely. It maintains close contact with the financial supervisory authority, the financial sector, and European as well as international partners to prepare for potential risks.
