US Financial Leaders Tackle New AI Cyber Threat

US Treasury Secretary Scott Bessent and Federal Reserve Chairman Jerome Powell recently called an urgent meeting with top Wall Street leaders. This was sparked by growing fears that a new artificial intelligence model from Anthropic could usher in a new era of cyber risks. The meeting took place last Tuesday at the Treasury Department in Washington. Its main goal was to ensure that banks are fully aware of the potential dangers linked to the "Mythos" model and similar AI systems, and that they are taking the necessary steps to strengthen their system protections. This hastily arranged and unannounced meeting highlights the increasing concern among regulators about the potential emergence of a new generation of cyberattacks, which could directly threaten the financial sector. The Mythos model is considered one of the most advanced systems. Anthropic indicates that it can detect vulnerabilities in major operating systems and browsers, and even exploit them when directed. This capability raises significant concerns about its potential use by hackers. These worries underscore Anthropic's cautious approach. They initially limited access to the model to a select group of major technology and finance companies, including Amazon, Apple, and JPMorgan. This was part of their "Glass Wing" project, which aims to secure critical systems before the widespread adoption of these models. The company also confirmed that it had been in communication with US officials before launching the model to discuss its offensive and defensive capabilities in the cybersecurity domain.