Anthropic To Brief Global Regulators On AI Cyber Risks
Anthropic is preparing to brief the Financial Stability Board on cyber vulnerabilities in the global financial system identified by its latest artificial intelligence model, according to a report by the Financial Times.
The startup behind the Claude chatbot will discuss the capabilities of its new Mythos Preview AI model with finance ministries and central banks linked to the FSB. The discussions follow a request from Andrew Bailey, the report said.
Bailey currently chairs the FSB, which coordinates financial regulations and risk oversight for G20 economies.
Mythos Model Raises Cybersecurity Concerns
Anthropic announced the Mythos model last month but has not yet released it publicly.
According to the company, the cybersecurity-focused AI system is designed to identify decades-old vulnerabilities across web browsers, infrastructure and software systems.
However, cybersecurity experts have warned that the technology could also enable more sophisticated cyberattacks. Concerns have intensified because many financial institutions still rely on ageing technology infrastructure and legacy systems.
As a result, regulators and central banks are increasingly examining how advanced AI models could affect financial stability and cybersecurity preparedness.
Regulators Monitor AI Threats Closely
An FSB spokesperson said the organisation welcomes engagement with Anthropic and other companies regarding emerging risks to global financial stability.
The planned briefing highlights growing concern among regulators over the rapid development of advanced artificial intelligence systems capable of identifying weaknesses in critical infrastructure.
Financial institutions worldwide have accelerated investments in cybersecurity protections as AI tools become more powerful and accessible.
Andrew Bailey Warned About Risks Earlier
In April, Bailey publicly warned that Mythos could create serious risks for global cybersecurity systems.
Speaking at an event at Columbia University in New York, Bailey said recent geopolitical tensions were no longer the only major concern facing the financial system.
He suggested the capabilities of Anthropic’s latest model could potentially expose vulnerabilities across a wide range of digital systems.
Bailey questioned whether the technology could identify weaknesses in other platforms that might later be exploited for cyberattack purposes.
The comments underscore growing fears among policymakers that advanced AI systems may outpace existing regulatory frameworks and cybersecurity protections.
With inputs from Reuters