Powell, Besant Warn Banks of Security Risks from Humanistic AI Myths: Bloomberg

US Treasury Secretary Scott Besent and Federal Reserve Chairman Jerome Powell reportedly held a meeting with Wall Street bank executives earlier this week to warn of cybersecurity risks associated with Anthropic’s new AI model.

According to A a report by BloombergThe meeting included executives from Citigroup, Bank of America, Wells Fargo, Morgan Stanley, and Goldman Sachs. Officials discussed Anthropic’s new Mythos AI model, which recently sparked widespread concern over its apparent advanced cybersecurity capabilities.

Officials held the meeting to ensure banks understand the risks posed by systems capable of identifying and exploiting software vulnerabilities across operating systems and web browsers, and to encourage institutions to strengthen defenses against potential AI-assisted cyberattacks targeting financial infrastructure.

Tools capable of automatically detecting vulnerabilities could lead to accelerated defensive security work and malicious hacking if misused, security researchers have warned.

Anthropic Mythos Model First appeared Online in March after draft materials about the system leaked online, revealing what the company described as its most capable AI model to date. During testing, the system reportedly found thousands of previously unknown software vulnerabilities, including zero-day flaws across major operating systems and web browsers.

The anthropological researchers at A a report Earlier this week, Mythos Preview’s vulnerability detection capabilities were not intentionally trained, but instead emerged through broader improvements to the model’s coding, logic, and independence.

“The same improvements that make the model significantly more effective at patching vulnerabilities also make it significantly more effective at exploiting them,” the company wrote.

Because of these capabilities, Anthropic has restricted access to a small group of cybersecurity organizations.

“Given the strength of its capabilities, we are examining how to launch it,” Anthropic said in a statement. “As is common practice across the industry, we are working with a small group of early access customers to test the model. We view this model as a step change and it is the most capable we have built to date.”

To address that riskAnthropic is testing Mythos through Project Glasswing, a collaboration with major technology and cybersecurity companies that uses the model to identify vulnerabilities in critical software and patch them before attackers can exploit them.

“The Glasswing project is the starting point. No organization can solve these cybersecurity problems alone,” the company said in a statement. statement. “Frontier AI developers, other software companies, security researchers, open source maintainers, and governments around the world all have essential roles to play.”

Anthropic did not respond immediately Decryption Request for comment.