Washington — The caput of the International Monetary Fund said she is acrophobic astir a almighty caller AI exemplary from Anthropic that poses large cybersecurity risks, informing "time is not our person connected this one."
Kristalina Georgieva, the IMF's managing director, said successful an interrogation acceptable to aerial Sunday connected "Face the Nation with Margaret Brennan" that the satellite does not person the quality "to support the planetary monetary strategy against monolithic cyber risks."
"The risks person been increasing exponentially," Georgieva said. "Yes, we are concerned. We are precise keen to spot much attraction to the guardrails that are indispensable to support fiscal stableness successful the satellite of AI."
On Tuesday, Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent held an urgent meeting with Wall Street leaders to sermon the cybersecurity risks posed by Anthropic's Claude Mythos Preview, sources told CBS News.
A Treasury Department spokesperson said successful a connection that "additional coordination meetings by Treasury are planned crossed a fig of regulators and institutions connected an ongoing ground to code these developments, arsenic good arsenic a big of different issues."
In a blog post earlier this week, Anthropic said the exemplary has demonstrated "a leap" successful the quality to spot cybersecurity vulnerabilities — immoderate of them decades aged — and exploit them. Anthropic is lone releasing the exemplary to prime partners truthful they tin usage it to harden their systems.
"Mythos Preview has already recovered thousands of high-severity vulnerabilities, including immoderate successful each large operating strategy and web browser. Given the complaint of AI progress, it volition not beryllium agelong earlier specified capabilities proliferate, perchance beyond actors who are committed to deploying them safely. The fallout — for economies, nationalist safety, and nationalist information — could beryllium severe," the institution said.
Georgieva said cardinal fiscal institutions, including cardinal banks, request to "work together" and beryllium "very attentive" successful managing the risks of cyberattacks.
"It is an contented that easy tin contiguous itself successful different parts of the world, and that is wherefore we request radical to cooperate," she said.
Watch much of the interrogation connected Sunday connected "Face the Nation with Margaret Brennan."
IMF main connected cybersecurity risks posed by AI
IMF main says she's acrophobic astir cybersecurity risks posed by Anthropic's latest AI model
(01:54)

2 hours ago
4




English (US) ·