Finance Minister Nirmala Sitharaman on Thursday held a high-level meeting with the heads of banks. In this meeting, potential threats to the banking sector from Anthropic’s ‘Claude Mythos’ AI model were discussed. This AI model is so advanced that hackers can use it to discover decades-old unknown vulnerabilities and launch cyberattacks on the financial sector. Banks instructed to remain on high alert The Finance Minister has instructed banks to take necessary steps to secure their IT systems and protect customer data. The Finance Ministry stated that the threat posed by Mythos is unprecedented. It requires better coordination and a high level of preparedness among financial institutions and banks. India’s plan to deal with the threat The Ministry of Finance is now planning to create a framework that can identify hacking attempts and take immediate action on them. Banks have been advised to create a real-time intelligence sharing mechanism to share information about cyber attacks among themselves. This will also include the Indian Computer Emergency Response Team (CERT-In) and other agencies. What is Claude Mythos and why is it dangerous? Mythos is Anthropic’s most powerful AI model. According to the company, Mythos has uncovered decades-old vulnerabilities in operating systems and browsers that humans could not see. Anthropic says that making it public would be like putting an advanced hacking tool in anyone’s hands. Anthropic has given access to only 40 selected companies like Amazon, Microsoft, and Google. However, according to reports, some unauthorized users have gained access to it. This raises fears that they could exploit the system’s security flaws. Anthropic stated that if it were released without any controls, the consequences for the economy, public safety, and national security could be very severe. AI broke out of its digital jail on its own Before the rollout of this model, an incident also came to light which shows how advanced it is. ‘Mythos’ was being tested in a secure ‘sandbox’ so that it could not use the internet. A sandbox can be called a kind of digital prison. But this AI broke through the security perimeter and made its own way out. This was discovered when a researcher working on it suddenly received an email sent by the same AI model. An AI breaking out on its own in this manner is extremely dangerous. High-level meeting on Mythos in America too Not only India, but also America is holding meetings on this issue. US Treasury Secretary Scott Bessent and Federal Reserve Chairman Jerome Powell recently held a closed-door meeting with senior bank officials. In it, they told bank officials to prepare their systems for potential threats arising from Mythos.
The post appeared first on .
