IndiaAI Safety Institute

The IndiaAI Safety Institute (IAISI) is India's national body for AI safety, incubated within the IndiaAI Mission under the Ministry of Electronics and Information Technology (MeitY). The institute was announced on January 30, 2025, by Minister Ashwini Vaishnaw, and formally launched by Prime Minister Narendra Modi on February 17, 2026, at the India AI Impact Summit held at Bharat Mandapam, New Delhi.

The institute's mandate is to ensure that AI in India is developed, deployed, and adopted in a safe, secure, inclusive, and trustworthy manner. It is designed to be a collaborative body rather than a regulatory one, working with academia, startups, industry, civil society, and government ministries to develop standards, frameworks, and tools tailored to India's unique social, economic, cultural, and linguistic diversity.

The IAISI operates on a hub-and-spoke model. Partner institutions — including IITs, national institutes of technology, autonomous research bodies, and private sector entities — each establish a dedicated IAISI Cell to carry out specific technical mandates. This distributed structure allows the institute to scale its research capacity across India while maintaining a central coordinating function.

Through Expressions of Interest, the institute onboarded eight initial Responsible AI projects in its first round, with over 400 applications received in a second round. A total of thirteen projects were active by the time of the February 2026 summit. Key research areas include machine unlearning (IIT Jodhpur), AI bias mitigation (NIT Raipur), privacy-preserving machine learning, explainable AI, AI auditing, and governance testing.

The institute sits under the Safe and Trusted AI pillar of the IndiaAI Mission, which has an allocated budget of Rs 20.46 crore. The broader IndiaAI Mission was approved by Cabinet in March 2024 with a Rs 10,372 crore, five-year mandate. Abhishek Singh, Additional Secretary at MeitY and CEO of IndiaAI, has been the key official overseeing the institute's development. The Safe and Trusted AI working group is co-chaired by Prof. Balaraman Ravindran of IIT Madras.

India used the February 2026 AI Impact Summit to position the IAISI as a model for the Global South, advocating for AI governance approaches that reflect developing-world contexts and championing a global compact on AI safety with 92 countries endorsing the New Delhi Declaration.