The German financial supervisory authority, BaFin, has warned about risks posed to banks by the new AI model “Mythos” from the US corporation Anthropic, as well as similar offerings. Speaking to the “Handelsblatt” (Wednesday edition), a BaFin spokesperson stated that since comparable approaches have already emerged, it is expected that other providers will also introduce such models. Consequently, financial institutions must anticipate that vulnerabilities could soon be discovered, necessitating quick and timely remediation.
Anthropic introduced its advanced model, Claude Mythos Preview, last week. This system has the capacity to independently and extensively identify security flaws within IT systems. Due to this capability, US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell briefly convened leading Wall Street representatives for a meeting to address Mythos. The goal was to raise awareness among risk institutes regarding the potential of this new AI model and similar systems, ensuring that appropriate protective measures are implemented.
According to financial circles, such high-level, ad-hoc meetings between banks and regulatory bodies like those held in the United States have not taken place in Germany or on the European level. Nevertheless, new cybersecurity threats are continuously discussed through ongoing exchanges between authorities and banks across Europe. It is reported that some major US banks, such as Goldman Sachs, are already utilizing “Mythos”.
However, German banks reportedly do not yet have access to the model. The Commerzbank and DZ Bank informed the newspaper that they are not currently using Mythos. The Deutsche Bank declined to comment on specific providers or AI models.
