Anthropic to brief global finance watchdog on AI found cyber flaws

In short: Anthropic will brief the Financial Stability Board about cyber security weaknesses its new AI model, Claude Mythos Preview, helped uncover.

What happened

Anthropic, a US AI company, has agreed to speak with members of the Financial Stability Board, according to two people familiar with the plan. The request came from Andrew Bailey, governor of the Bank of England, who chairs the group.

The Financial Stability Board, or FSB, is a global watchdog that brings together finance ministries, central banks, and market regulators from G20 countries. It includes officials from places such as the US, UK, Canada, France, Germany, Japan, Saudi Arabia, Australia, and China.

Anthropic said last month that its Claude Mythos Preview model found “thousands of high-severity vulnerabilities” across major operating systems and web browsers. A vulnerability is a weakness in software that attackers can use, like a broken lock on a door. Anthropic has limited access to Mythos to about 40 organisations, including Amazon, Microsoft, and JPMorgan Chase, and it agreed not to distribute it more widely after a request from the White House.

Why it matters

Banks and payment systems depend on software working safely and reliably. If AI tools make it easier to find weak spots, attackers could hit more targets faster, including in countries that have fewer resources to defend themselves. Regulators are already pushing banks to patch software more quickly, and the FSB is preparing guidance on how to use AI safely in the financial system.

Source: Financial Times