Powell, Bessent Warn Banks About Safety Dangers From Anthropic's Mythos AI: Bloomberg – Decrypt

U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell reportedly convened a gathering with Wall Road financial institution CEOs earlier this week to warn about cybersecurity dangers tied to a brand new synthetic intelligence mannequin from Anthropic.

Based on a report by Bloomberg, the assembly included executives from Citigroup, Financial institution of America, Wells Fargo, Morgan Stanley, and Goldman Sachs. Officers mentioned Anthropic’s new AI mannequin Mythos, which has lately drawn broad concern over its obvious superior cybersecurity capabilities.

Officers convened the assembly to make sure banks perceive the dangers posed by techniques able to figuring out and exploiting software program vulnerabilities throughout working techniques and internet browsers, and to encourage establishments to strengthen defenses towards potential AI-assisted cyberattacks concentrating on monetary infrastructure.

Safety researchers have warned that instruments able to robotically discovering vulnerabilities may speed up each defensive safety work and malicious hacking if misused.

Anthropic’s Mythos mannequin first surfaced on-line in March after draft supplies in regards to the system leaked on-line, revealing what the corporate described as its most succesful AI mannequin but. In testing, the system reportedly discovered 1000’s of beforehand unknown software program vulnerabilities, together with zero-day flaws throughout main working techniques and internet browsers.

Anthropic researchers stated in a report earlier this week that Mythos Preview’s vulnerability-discovery capabilities weren’t deliberately skilled, however as an alternative emerged from broader enhancements within the mannequin’s coding, reasoning, and autonomy.

“The identical enhancements that make the mannequin considerably simpler at patching vulnerabilities additionally make it considerably simpler at exploiting them,” the agency wrote.

Due to these capabilities, Anthropic has restricted entry to a small group of cybersecurity organizations.

“Given the power of its capabilities, we’re being deliberate about how we launch it,” Anthropic stated in an announcement. “As is customary follow throughout the trade, we’re working with a small group of early entry clients to check the mannequin. We contemplate this mannequin a step change and essentially the most succesful we’ve constructed thus far.”

To deal with that danger, Anthropic is testing Mythos via Mission Glasswing, a collaboration with main expertise and cybersecurity corporations that makes use of the mannequin to establish and patch vulnerabilities in important software program earlier than attackers can exploit them.

“Mission Glasswing is a place to begin. Nobody group can resolve these cybersecurity issues alone,” the corporate stated in an announcement. “Frontier AI builders, different software program corporations, safety researchers, open-source maintainers, and governments internationally all have important roles to play.”

Anthropic didn’t instantly reply to Decrypt’s request for remark.