Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell summoned financial institution executives for a gathering this week the place they inspired the executives to make use of Anthropic’s new Mythos mannequin to detect vulnerabilities, according to Bloomberg.
Certainly, whereas JPMorgan Chase was the one financial institution listed as one of many preliminary associate organizations with entry to the mannequin, Goldman Sachs, Citigroup, Financial institution of America, and Morgan Stanley are reportedly testing Mythos as properly.
Anthropic introduced the mannequin this week however stated it will be limiting entry for now, partially as a result of Mythos — regardless of not being skilled particularly for cybersecurity — is simply too good at discovering safety vulnerabilities. (Others instructed this was hype or just a sensible enterprise gross sales technique.)
The report is especially shocking since Anthropic is at the moment battling the Trump administration in court docket over the Division of Protection’s designation of Anthropic as a supply-chain threat; that designation got here after negotiations fell aside over the corporate’s efforts to restrict how its AI fashions can be utilized by the federal government.
In the meantime, the Financial Times reports that U.Ok. monetary regulators are additionally discussing the chance posed by Mythos.
