Banks Scramble to Access Mythos AI

The emergence of Anthropic’s Mythos AI model is prompting a scramble among global banks to gain access and test the technology, even as regulators rush to assess the cybersecurity risks the model poses, as reported by Reuters.

Access to Mythos has so far been restricted to partners in Anthropic’s Project Glasswing initiative and around 40 additional organizations. JPMorgan is the only bank Anthropic itself has publicly confirmed has access, though Bank of America, Goldman Sachs, Morgan Stanley, and Citigroup also reportedly have access to the model. Multiple senior banking and regulatory sources in Europe told Reuters they were not aware of any European financial institution that had been granted access, though recent reporting suggests Anthropic plans to do so.

Several European supervisory sources told Reuters they are not yet overly concerned and are monitoring developments through existing cyber resilience frameworks. However, the European Central Bank and other regulators have reportedly been in contact with European banks to assess their preparedness and awareness of the threat. The British government wrote to Anthropic on April 15 stating that testing by its AI Security Institute found Mythos to be “substantially more capable at cyber offence than any model we have previously assessed.”

Deutsche Bank CEO Christian Sewing struck a measured tone. “It’s certainly not something that’s causing panic or setting off any alarm bells on our end right now, but it’s definitely something we need to keep in mind in our day-to-day risk management,” he told journalists. Barclays CEO C. S. Venkatakrishnan was more pointed, describing Mythos as a serious threat to the global banking system and warning it was likely to be followed by similar, more powerful cyberthreats.