Anthropic to share Mythos cyber flaw findings with global finance watchdog | AI (artificial intelligence)
Publish Date: 2026-05-18 07:20:00
Source Domain: www.theguardian.com
Here’s a polite and respectful summary of the article using an unordered list:
– Anthropic, the US startup, is preparing to brief the global finance watchdog about its advanced AI model, Claude Mythos, highlighting concerns about its potential impact on cyber defences.
– Mythos has alarmed experts due to its ability to uncover previously unknown flaws in IT systems, which could be exploited by hackers.
– To mitigate risks, Anthropic has restricted the public release of Mythos and instead granted access to a select group of tech companies and banks, including Apple and JP Morgan.
– The Financial Stability Board, chaired by Andrew Bailey, will discuss the implications of Mythos to monitor and manage risks in the global financial system.
– The UK’s AI Security Institute reported a “notable capability jump” in the latest version of Mythos, which successfully completed a tough cybersecurity test for the first time.
– The Financial Stability Board monitors global financial stability and its inclusion of cyber risks from AI aligns with broader international concerns, such as those raised by the International Monetary Fund.
– While some experts caution against overstating Mythos’ threat, they acknowledge the evolving nature of cyber threats and the necessity for enhanced oversight and testing.
