Anthropic has agreed to brief leading finance ministries and central banks on cyber security weaknesses in the global financial system identified by its latest AI model, the Financial Times reported citing unnamed sources.
The briefings are said to follow a request from Bank of England governor Andrew Bailey, who asked the US technology company to present the capabilities of its Claude Mythos Preview model to members of the Financial Stability Board (FSB), which he chairs.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
The FSB is an international body that brings together finance ministry officials, central bankers and securities regulators from G20 economies. Its membership includes officials from the US, UK, Canada, France, Germany, Japan, Saudi Arabia, Australia and China.
The report said a growing number of FSB members are concerned that Mythos, as well as AI systems developed by other US technology groups, could expose weaknesses in banks’ cyber defences and create broader risks for the global financial system.
Anthropic said last month that Mythos had “found thousands of high-severity vulnerabilities, including some in every major operating system and web browser”.
It added: “The fallout – for economies, public safety and national security – could be severe.”
Access to Mythos has been restricted to a small number of organisations, largely in the US, because of concerns about the consequences if the model were misused.
That limited availability has also raised concerns among regulators and companies outside the US about uneven levels of protection.
The company has received many requests from other countries for access to the model or for briefings on its capabilities.
It has agreed to provide high-level briefings to some non-US institutions, including the European Commission.
About 40 organisations have been given access to Mythos, including Amazon, Microsoft and JPMorgan Chase, so they can address the vulnerabilities it identifies.
However, the report said Anthropic has agreed not to distribute the model more widely after a request from the White House.
Alongside this, the FSB is preparing a report setting out “sound practices” for the use of AI in the financial system. It plans to release the document for consultation next month.
Neither Anthropic nor the FSB commented on the reported communication between them.
Regulators have meanwhile been urging banks and other financial institutions to review their cyber security arrangements and accelerate software patching in response to vulnerabilities identified by new AI models.
Earlier this month, the IMF urged policymakers to strengthen international co-operation to address the cyber security vulnerabilities highlighted by the latest AI models.
The organisation warned that such systems “elevate cyber risk to a potential macro-financial shock”. “Cyber risk does not respect borders,”
IMF officials wrote in a blog post. “Emerging and developing countries, which often have more severe resource constraints, may be disproportionately exposed to attackers targeting regions with weaker defences.”
