Indian government, tech firm join forces to test Anthropic’s Claude Mythos: Report

Anthropic’s new cybersecurity-focused model, Claude Mythos, has sparked global concerns across nations and industries over its capabilities and potential misuse. Following a month of testing by partner organisations under Project Glasswing, the Indian government and tech firms have now joined forces to test the model and assess vulnerabilities across its critical systems.

India to test Mythos Preview

According to a Bloomberg report, several Indian tech giants, including Infosys Ltd. and Tata Consultancy Services Ltd. (TCS), will likely get access to Mythos Preview. These companies will reportedly test their software systems to identify potential threats, weaknesses or security vulnerabilities. It was highlighted that Infosys will be testing Mythos capabilities for creating security fixes for its widely used banking platform, Finacle.

Must read: Anthropic Project Glasswing: Mythos Preview flagged over 10,000 security flaws across critical systems

India’s government cybersecurity agency, CERT-In, will also be testing critical software systems, including the Aadhaar national ID infrastructure and government login platforms, to identify possible security weaknesses. However, the report also highlighted that many companies do not have direct access to Mythos. Instead, they are using the Claude Opus 4.7 model to help identify and patch software vulnerabilities.

Previously, the Mythos access was limited to over 50 companies and governments. Anthropic’s recent finding revealed that the model has identified over 23,000 vulnerabilities and over 10,000 critical flaws within the initial months of testing by organisations like Cloudflare, Mozilla, and others. 

India plays a backbone role in global banking and technology operations. Furthermore, the country has rapidly expanded its digital infrastructure, including large-scale digital initiatives such as Aadhaar, the national ID system that digitally registers all 1.4 billion citizens. Therefore, the need to proactively identify and address cyber vulnerabilities has become crucial.

Must read: Anthropic’s Mythos AI raises banking cyber risk concerns, FSB briefing planned

How does India plan to test Mythos?

The report revealed that India is reportedly negotiating access to Mythos in a secure setup with US officials. The government is said to have designated a testing facility and identified relevant sectors and specialists to assess the model’s performance and risks.

Bloomberg quoted US Ambassador Sergio Gor, “They (India) have requested access to various things, and as a trusted partner, those are some things that we are definitely looking at.” Therefore, talks are expected to continue as both sides work toward establishing a secure and mutually agreed framework for evaluating Mythos.

Must read: BT explainer: Inside Claude Mythos, the AI is forcing a rethink of global cybersecurity

Cybersecurity preparedness

Since the launch of Mythos Preview, governments around the world have started to advise banking and financial companies to strengthen their cybersecurity frameworks, conduct regular vulnerability assessments, and bring stricter safeguards to mitigate potential risks associated with AI models.

In April, India’s Finance Minister Nirmala Sitharaman also held a meeting with Indian banks to warn them about potential threats from AI models like Mythos. It was highlighted that the government is closely monitoring the situation, and has instructed banks to increase vigilance to protect IT systems, customer data, and financial assets. Now with goining tests for Mythos, it may give the country an understanding of its potential cyberseucirty risks and capabilities, helping strengthen cybersecurity preparedness.