The second day of the India AI Impact Summit 2026 featured a key session titled ‘AI in Governance: Revolutionising Government Efficiency’, where global researchers and senior policymakers discussed how AI can enhance public service delivery at scale. The discussion focused on moving beyond pilot projects to measurable impact, with emphasis on rigorous evaluation, responsible deployment and systems-level readiness. Social Entrepreneur Dean Karlan presented research on the use of machine learning to improve targeting of public services in Togo. 

 

The study showed measurable gains in food security, mental health and socio-economic indicators when AI-supported targeting was applied. However, it also highlighted limitations, noting that phone metadata alone failed to capture treatment effects and revealed risks of model drift in predicting short-term vulnerability. Speakers stressed the importance of experimentation, iterative testing and evidence-based AI procurement before large-scale deployment, especially in beneficiary identification and resource allocation. The panel discussion assessed India’s preparedness for AI in governance.