RBI in talks with banks to understand the risks posed by Anthropic's Mythos

India's central bank is in talks with global regulators, Indian lenders and government officials to understand the potential risks posed by Anthropic's new artificial intelligence model Mythos. Just like global regulators, the Reserve Bank of India's preliminary assessment suggests Mythos could pose cybersecurity risks by accelerating the discovery and exploitation of software vulnerabilities. Regulators in Asia, Europe and the United States have already warned banks to review defences and preparedness.

In a recent development, India Finance Minister Nirmala Sitharaman summoned a high-level with heads of banks to review emerging cybersecurity risks associated with advanced artificial intelligence models.   

During the meeting, Sitharaman directed banks to take proactive steps to strengthen their IT systems, protect customer data, and secure financial assets.

“It was advised that a robust mechanism for real-time threat intelligence sharing may be established among banks, @IndianCERT and other relevant agencies so that emerging threats are identified early and disseminated across the ecosystem without delay,” the finance ministry said in a post on X.

Banks were also advised to promptly report any suspicious activity or cyber incidents to authorities such as CERT-In and maintain close coordination with all concerned agencies, as per the post.  

The recommendations were provided during the meeting that focused on assessing the risks posed by recent advancements in AI, particularly the possibility of such technologies being misused to exploit software vulnerabilities.

Sitharaman further urged the Indian Banks' Association to establish a coordinated institutional response mechanism for dealing with such threats. She also called on banks to engage top cybersecurity experts and specialized agencies to continuously enhance their defense and monitoring capabilities.

While acknowledging the progress made by banks in strengthening cybersecurity, she highlighted that the risks posed by next-generation AI models are unprecedented, requiring heightened vigilance, preparedness, and collaboration across the financial ecosystem.

A senior Finance Ministry official noted that both the ministry and the Reserve Bank of India are currently assessing the extent of potential risks to the financial sector. He added that Indian systems remain secure, and there is no immediate cause for concern, with the RBI conducting its own due diligence.

As per the reports, Anthropic said Mythos can outperform humans at cyber-security tasks, finding and exploiting thousands of bugs, including 27-year-old vulnerabilities, in major operating systems and web browsers.

Anthropic, an U.S.-based artificial intelligence company, said unauthorised access was made on its new model Mythos, which is deemed too dangerous for public release.

Announced on April 7, Mythos is being deployed as part of Anthropic’s ‘Project Glasswing’, a controlled initiative under which select organisations “are permitted to use the unreleased Claude Mythos Preview model for defensive cybersecurity”.

Mythos is a powerful AI model that has sparked concerns among regulators about its unprecedented ability “to identify digital security vulnerabilities and potential for misuse”.

Anthropic chose not to release Mythos publicly, arguing that its capabilities pose unprecedented cybersecurity risks, as per reports.