Anthropic myth sends banks rushing to address cyber risks

The country’s major banks are increasingly concerned about the ability of next-generation artificial intelligence models like Anthropic’s Mythos to compromise their systems and say they are stepping up efforts to head off the risk of a serious compromise to cybersecurity defenses.

Anthropic said earlier this month that it would not release Mythos publicly, citing cybersecurity risks as the model is capable of detecting unknown defense vulnerabilities, putting the entire global financial system at risk.

Load…