IndiaAI Mission funds 13 trusted AI projects on deepfakes and bias

India’s Safe and Trusted AI update shows what a middle path between pure industrial policy and pure risk aversion can look like. Rather than treating safety as a brake on deployment, MeitY is funding deepfake detection, bias mitigation and privacy preserving learning alongside a portfolio of domestic LLMs and small language models.([ddindia.co.in](https://ddindia.co.in/2026/07/indiaai-missions-safe-trusted-ai-pillar-advances-responsible-ai-13-projects-approved-to-tackle-deepfakes-and-bias/)) That combination matters in the AGI race because it increases India’s capability surface area while simultaneously building its own expertise in the parts of the stack that will attract the most regulatory scrutiny over the next decade.

The numbers are not trivial: 20 sovereign model proposals, 15 empanelled compute providers, tens of lakhs of GPU hours and dozens of AI labs and Centres of Excellence.([ddindia.co.in](https://ddindia.co.in/2026/07/indiaai-missions-safe-trusted-ai-pillar-advances-responsible-ai-13-projects-approved-to-tackle-deepfakes-and-bias/)) India is effectively building a parallel, state backed ecosystem that does not depend entirely on US or Chinese APIs, especially for sensitive sectors like media, healthcare and public services. For other emerging economies, this is an influential template: treat AI safety institutes, governance guidelines and technical mitigations as part of the same mission as scaling compute. In the long run, that could make India a key swing player in global AGI governance, with enough indigenous capability to negotiate from strength rather than fear.