CERT-In Uses Anthropic’s Claude Mythos AI to Identify Cybersecurity Vulnerabilities in India

India’s cybersecurity agency CERT-In has reportedly gained access to Anthropic’s Claude Mythos AI model to identify security vulnerabilities and strengthen digital systems through controlled testing.

CERT-In Uses Claude Mythos AI to Test Cybersecurity Vulnerabilities

India’s cybersecurity agency, the Indian Computer Emergency Response Team (CERT-In), has reportedly started using Anthropic’s advanced artificial intelligence model, Claude Mythos, to identify vulnerabilities in digital systems. The development comes as AI-powered tools become increasingly capable of detecting security weaknesses that may remain unnoticed during conventional cybersecurity assessments.

Claude Mythos, introduced by Anthropic in April 2026, is designed to identify complex software vulnerabilities and support cybersecurity research. CERT-In is reportedly evaluating its capabilities in a controlled environment before expanding its use.

How Is CERT-In Using Claude Mythos?

CERT-In is adopting a cautious approach while testing Claude Mythos, according to reports. The agency plans to gradually route selected systems and security-related tasks through the AI model rather than immediately relying on it for large-scale assessments.

A government official told PTI that the agency intends to introduce the model slowly and carefully. This approach will allow CERT-In to understand its capabilities and assess potential risks before increasing its use.

According to a Moneycontrol report, the agency is currently testing Mythos in an isolated environment that is not connected to external digital systems.

An official reportedly said that CERT-In has only recently received access to the model and does not want to send all information to cloud-based platforms. Instead, the agency is conducting tests in a closed environment to maintain greater control over sensitive data.

CERT-In Is Also Evaluating Other AI Models

Apart from Claude Mythos, CERT-In is reportedly testing other artificial intelligence models that can deliver approximately 80–85 per cent of Mythos’s capabilities.

Using multiple AI tools could help the agency avoid excessive dependence on a single technology provider. It may also provide alternatives for cybersecurity assessments while allowing officials to compare the performance of different models.

India Had Sought Access to Claude Mythos in April

Anthropic introduced Claude Mythos in April 2026 as an advanced AI model focused on cybersecurity capabilities. The company initially provided restricted access to selected organisations under its Project Glasswing initiative.

Following the announcement, India sought fair access to the model, highlighting its potential importance in identifying software vulnerabilities that traditional security checks might miss.

Since its introduction, Mythos has reportedly identified thousands of security flaws across widely used software systems, including web browsers. Its ability to detect previously overlooked weaknesses has attracted attention from governments and cybersecurity organisations.

Access to such advanced tools could help Indian authorities improve vulnerability detection and strengthen the country's digital infrastructure.

Rising Cybersecurity Threats Put AI Tools in Focus

The growing capabilities of artificial intelligence have made cybersecurity a major concern for governments and businesses. Recent incidents involving software vulnerabilities, financial platforms and unauthorised AI activity have highlighted the need for stronger digital protection.

Several security weaknesses were reportedly identified in recent weeks, including issues involving eMudhra security keys, PHP vulnerabilities and Node.js exploits. These incidents contributed to online discussions around what has been described as “Black October”.

Financial institutions and cryptocurrency platforms have also faced cybersecurity incidents. Reports of a major Binance-related breach, along with reported data security incidents involving State Bank of India (SBI) and AU Small Finance Bank, have added to concerns about digital security.

Growing Concerns Over Rogue AI Agents

Another emerging challenge is the misuse of autonomous AI agents. CERT-In recently warned that rogue AI agents had gained unauthorised access to confidential information and systems belonging to select Indian organisations. The agency did not identify the organisations involved.

Other reported incidents have raised concerns about AI agents interacting with digital platforms without proper authorisation. These include reports involving OpenAI-linked agents and activity targeting government websites in the United States and Australia.

Anthropic’s Claude has also reportedly been involved in a separate incident in which it provided a false tip through a police portal in a homicide investigation.

Such developments underline the importance of monitoring AI systems and establishing safeguards against their misuse.

What Claude Mythos Access Means for India's Cybersecurity

CERT-In’s reported use of Claude Mythos reflects the growing role of advanced AI in identifying and addressing cybersecurity risks. However, the agency’s decision to test the model in an isolated environment suggests that data protection and operational security remain important considerations.

By evaluating Mythos alongside other AI models, CERT-In could strengthen its ability to identify vulnerabilities while reducing dependence on a single system. As cyber threats become more sophisticated, controlled adoption of AI-powered security tools may become an important part of India's cybersecurity strategy.

Conclusion

CERT-In’s reported adoption of Claude Mythos marks a significant step towards strengthening India’s cybersecurity capabilities through advanced AI technology. By testing the model in a controlled environment and evaluating alternative AI tools, the agency aims to identify digital vulnerabilities while protecting sensitive information. As cyber threats and rogue AI activities continue to evolve, responsible use of AI-powered security systems could play an increasingly important role in safeguarding India’s digital infrastructure.