India is conducting high-level security testing on some of its most sensitive government and financial software systems to assess vulnerabilities against Anthropic’s next-generation AI model Mythos, according to officials familiar with the matter.
Major Indian IT firms Infosys Ltd and Tata Consultancy Services Ltd are reportedly involved in running controlled tests to identify weaknesses in widely used platforms, including banking and government systems, the officials said.
Infosys is said to be focusing in particular on strengthening its Finacle banking software, while India’s state cybersecurity agency CERT-In is also reviewing critical infrastructure such as the Aadhaar national ID system and government login services.
Although the companies do not currently have direct access to Mythos, they are using Anthropic’s Claude Opus model in a secure environment to help detect and patch vulnerabilities, officials added.
Concerns over the AI system are growing globally, with governments and corporations increasingly cautious about its potential to both identify and enable sophisticated cyberattacks.
India, which hosts some of the world’s largest outsourcing and back-office technology providers, has placed added urgency on strengthening cybersecurity as its digital public infrastructure continues to expand rapidly.
Authorities say discussions are ongoing with the United States regarding possible secure access to Mythos for testing purposes, with India seeking a controlled domestic environment for evaluation.
Officials also noted that regulatory bodies, banks, and government agencies are increasing vigilance as AI-driven cybersecurity risks become a growing global concern.
