German banks examine risks of Anthropic's AI model Mythos | News.az

The model, known as Mythos, has drawn attention from cybersecurity experts who warn it could pose challenges for the banking sector, particularly for institutions relying on legacy systems, News.Az reports, citing Reuters.

According to Kolja Gabriel of the German Banking Association, discussions are underway with cyber specialists at member banks as well as with Germany's finance ministry and other authorities.

Regulators warned that financial institutions must be ready to respond quickly if new security weaknesses are identified. Meanwhile, IT security firms are testing the model in controlled environments to help identify and fix potential risks.

Anthropic has said its current version, Claude Mythos Preview, will not be publicly released and is instead being evaluated privately by selected organizations, including major tech firms and JPMorgan Chase, to strengthen defenses against emerging threats.