Finance Minister Nirmala Sitharaman has raised the alarm over what she calls an 'unprecedented emerging threat' from Anthropic's Claude Mythos AI to banks and the broader financial system. According to reports from The Hindu, Sitharaman made this announcement after chairing a high-level meeting with senior bankers and regulatory officials to assess potential cybersecurity threats posed by advanced AI models. The minister emphasized that banks will need to do more to protect consumers from these evolving threats. As per the Finance Ministry's official communication, the threat was serious enough to require 'a very high degree of vigilance, preparedness and better coordination across financial institutions'. The response was telling: Claude Mythos wasn't just another routine launch, but a possibility that AI systems were becoming good enough to find serious flaws in critical software much faster than humans.
SBI Chairman Challa Srinivasulu Setty will lead the efforts to counter the AI-borne cybersecurity challenge. As reported by The Economic Times, Sitharaman held a meeting on Thursday with representatives of the Indian banking system regarding the emerging threat posed by Claude Mythos AI. The minister indicated that banks will discuss investments, technologies and the use of AI to tackle this emerging threat, with significant interaction expected in the coming weeks. According to Forbes India, the meeting brought together bank chiefs, RBI officials and representatives of the Ministry of Electronics and Information Technology for an unusual reason: A new AI model that governments fear could change the balance between cyber attackers and defenders. The government is now pushing for real-time threat intelligence-sharing system among banks and agencies like CERT-In, with officials stressing that while there's no immediate breach reported in India, the risk level is serious enough to warrant urgent preparation.
According to The Economic Times, Sitharaman stated that banks will explore investments and technologies, including AI, to counter this new AI-borne challenge. The minister emphasized that banks will need something new, something far more versatile to be able to counter the newer threats which are likely to come. She indicated that in the coming weeks, there will be a lot of interaction within the banks about understanding where more investment will have to be made, what kind of technologies can come in and how AI itself can be used for countering this AI-borne challenge. As per Forbes India, the meeting focused on preparedness, coordination and real-time threat sharing, with banks asked to strengthen defences, work more closely with cybersecurity experts and build stronger information-sharing channels with CERT-In and other agencies. The Ministry of Finance confirmed that banks have made progress in strengthening cybersecurity systems but warned that emerging AI-driven threats require enhanced cyber monitoring frameworks and engagement with specialized cybersecurity experts.
Claude Mythos Preview, announced by Anthropic on April 7, is part of Project Glasswing, a restricted cybersecurity initiative involving companies including Amazon, Apple, Google, Microsoft and Cisco. According to Forbes India, Anthropic says the model can identify zero-day vulnerabilities, or previously unknown software flaws, across major operating systems and web browsers, and can also chain bugs together into more complex exploits. The examples cited are striking: a 27-year-old bug in OpenBSD, an operating system known for its security, and a 16-year-old flaw in FFmpeg, a widely used media-processing tool that sits underneath countless apps and services. In one pre-release test, an earlier version of Mythos bypassed sandbox restrictions and found ways to gain broader internet access, contacting a researcher running the evaluation. As per Forbes India, some testing has been done in limited environments with a lot of human direction, though AI tools are genuinely impressive from a security perspective. Even Anthropic itself has reportedly restricted public release of Mythos due to these risks, highlighting the seriousness of the threat.
The threat extends beyond India to regulators and banks worldwide, from India to Europe and beyond, who are monitoring Mythos closely. Financial leaders warn it could outpace current cyber defences, especially in systems relying on older technology. Some reports even suggest unauthorised access to the model, raising fears it could already be in the wrong hands. This situation highlights a new phase in AI risk: AI is no longer just about automation or productivity. It's now capable of finding and exploiting weaknesses faster than humans can fix them, creating a cybersecurity arms race between defenders and attackers. Sitharaman's warning isn't about a specific attack, it's about 'capability risk'. Tools like Mythos could fundamentally change how cyberattacks happen, especially in highly interconnected sectors like banking. The longer-term challenge is that powerful cyber capability may not remain confined to frontier labs or large vendors for very long, as Yash Kadakia, founder of Security Brigade, explains: "Think of it like this: Earlier, we were fighting wars with guns. Now everyone will eventually have access to tanks."
