India Asks Sarvam AI, BharatGen to Develop Cyber AI Models

India's government has reportedly requested homegrown AI companies Sarvam AI and BharatGen to develop advanced cybersecurity capabilities similar to Anthropic's Mythos model. This initiative is part of India's broader strategy to decrease its dependence on foreign frontier AI models for securing critical infrastructure.

The developed models are intended to operate on the government's isolated compute infrastructure and will be deployed to safeguard critical systems against cyberattacks. However, a specific timeline for the development of these capabilities has not yet been provided.

Currently, the Indian government utilizes a mix of open-source and indigenous AI models, including Sarvam, to identify and address security vulnerabilities within its critical infrastructure.

India's national cybersecurity agency, CERT-In, has been evaluating alternative AI models in a controlled environment, reporting that current models offer approximately 60-70% of the capabilities found in models like Mythos. This evaluation helps in probing code and identifying security gaps.

The push to develop these indigenous capabilities reflects a wider national objective to enhance sovereign cybersecurity defenses and reduce reliance on external AI providers for protecting essential national assets.