Anthropic is set to brief the global financial watchdog on the implications of its Claude Mythos AI model, which experts say could pose risks to cyber defence systems.

The US-based AI company will discuss Mythos with the Financial Stability Board, chaired by Bank of England Governor Andrew Bailey.

Anthropic has not released the model publicly, after saying it has advanced capabilities that can detect previously unknown vulnerabilities in IT systems that could potentially be exploited by hackers.

The company has instead provided access to Mythos to a select group of technology firms and financial institutions, including Apple and JPMorgan, to help them test for potential security weaknesses the model might identify.

An FSB spokesperson confirmed the arrangement, saying the regulator “welcomes engagement with Anthropic and other firms on emerging and frontier risks to global stability.”

The UK’s AI Security Institute, which evaluates advanced AI systems, released an updated assessment of Mythos last week after examining the version shared with banks and tech companies.

It said the latest version it reviewed represented a “notable capability jump” compared with the preview it tested the previous month.

“Frontier AI’s autonomous cyber and software capability is advancing quickly: the length of cyber tasks that frontier models can complete autonomously has doubled on the order of months, not years,” the AISI said.

It said it is developing more advanced cybersecurity tests to better track the progress of AI models.

The FSB, which monitors and issues recommendations on the global financial system, brings together officials from major economies including the US, UK, Australia and China.

Its steering committee comprises senior central bank and finance ministry representatives.

Earlier this month, the International Monetary Fund warned that financial stability risks are increasing due to “fast-moving” advances in artificial intelligence, calling for a coordinated global response.