The Ministry of Electronics and Information Technology (MeitY) has told government ministries to pause deploying OpenAI and Anthropic AI models for cybersecurity. This advisory lands just days after both companies reportedly pitched their tools directly to multiple Indian government departments. It signals a deepening government concern over data security and foreign AI in critical national functions.
How We Got Here
This cautious stance isn't new; the Finance Ministry had warned against ChatGPT use on office devices in January 2025 due to confidentiality risks. The advisory also follows a June 29 Moneycontrol report that several government organizations had already restricted employees from using unapproved external AI platforms.
The Numbers
- The MeitY office memorandum specifically targets AI models for "cybersecurity and related functions."
- This follows a recent CERT-In warning that advanced AI models can enable more sophisticated cyberattacks.
- CERT-In also advised government organizations to patch critical vulnerabilities within 12 to 24 hours.
- In November 2025, India's AI Governance Guidelines adopted a risk-based approach, relying on existing laws, not new ones.
- The June advisory mandates using only approved AI platforms for sensitive government work, not a blanket ban.
What Happens Next
🇮🇳 Why This Matters for India
For Indian AI startups developing solutions for critical sectors like defense or finance, this creates a clear opening to differentiate on data sovereignty and trust over foreign models.
The Take
This clearly signals the government's strong preference for a "Made in India" AI stack for critical functions. Expect indigenous AI players to get a significant tailwind in sensitive government contracts over the next 12-18 months.
Source:
MediaNama ↗