India is testing critical digital and banking systems against vulnerabilities linked to Anthropic’s AI tools.
Infosys, TCS and CERT-In are among entities involved in the assessments.
The move comes amid growing global concerns over the cybersecurity risks posed by advanced AI models.
India is testing some of its most sensitive public-facing financial and government software applications to better understand possible vulnerabilities to Anthropic’s next-generation AI model, Mythos, Bloomberg reported, citing Indian officials familiar with the matter.
The tests are being conducted by major Indian technology firms including Infosys and Tata Consultancy Services in a secure environment. Infosys is also working on identifying and patching vulnerabilities in its widely used Finacle banking software, the report said.
1 May 2026
Get the latest issue of Outlook Business
Separately, India’s cybersecurity agency CERT-In is testing key digital infrastructure such as the Aadhaar national ID programme and government login systems. Since access to Mythos itself remains restricted, organisations are reportedly using Anthropic’s Claude Opus 4.7 AI model to identify and fix vulnerabilities.
Global AI Concerns
The assessments reflect growing concerns among governments and businesses worldwide over the cybersecurity risks linked to advanced AI systems.
Mythos, developed by Anthropic, is designed to detect software vulnerabilities but has also raised fears because of its potential ability to power sophisticated cyberattacks. Access to the model has so far been limited to a small group of organisations including Apple and JPMorgan Chase under Anthropic’s Project Glasswing initiative.
India’s urgency is tied to both its large software services industry and the rapid expansion of government-backed digital infrastructure under Prime Minister Narendra Modi. Indian IT firms manage critical systems for banks, businesses and government services globally.
Infosys operates India’s service-tax platform, while TCS manages the country’s passport system. Finacle, developed by Infosys, is used by financial institutions in several countries.
India Seeks Wider Access
Bloomberg reported that India’s central bank has already held multiple meetings with banks to discuss potential threats linked to Mythos and strengthen cybersecurity preparedness.
Last month, Finance Minister Nirmala Sitharaman urged banks to increase vigilance around IT systems, customer data and financial resources amid concerns over emerging AI-driven cyber risks.
Infosys CEO Salil Parekh had also said recently that Mythos was exposing more vulnerabilities than one thought possible previously while adding that the technology could create new business opportunities around cybersecurity solutions.
India is now reportedly seeking access to Mythos in a secure environment within the country, with discussions involving the Ministry of External Affairs and US officials. US Ambassador to India Sergio Gor recently said to Bloomberg that Washington was reviewing some of India’s requests as part of broader cooperation discussions.
