J-PAL at the India AI Impact Summit 2026
India hosted its flagship international gathering on AI, the India AI Impact Summit 2026, from February 16-20, 2026 in New Delhi. The summit was special for two reasons. It was the first of its kind to be held in the Global South, and it was the first to place impact at the heart of conversations on AI innovations.
The J-PAL team had a whirlwind week at the summit, organizing our official Summit event, AI for Social Good: Impact That Works at Bharat Mandapam in New Delhi on February 17, 2026 as well as speaking at events organized by Data.org, Open AI and the World Bank.
In partnership with
| 
| 
| 
|
AI for Social Good: Impact That Works
As a fast-improving, general-purpose technology accessible to anyone with a smartphone, AI is already having profound effects on our lives. The excitement around AI, as justified as it is, has also brought questions on effectiveness, transparency, and fairness into sharp focus, especially when AI tools are designed for underserved communities.
J-PAL's official event at the India AI Impact Summit 2026 — AI for Social Good: Impact That Works — was a day-long research and policy seminar held on February 17, 2026 at Bharat Mandapam in New Delhi to examine these questions and chart a path toward an AI ecosystem that maximizes social good by placing rigorous evaluations at the core.
The seminar featured senior government officials leading India's national AI strategy, researchers evaluating cutting-edge AI applications, and leaders from global tech companies, nonprofits, and key funders. Together, they dived into the vision for catalyzing AI to advance social welfare and what must change for AI applications to deliver sustained impact at scale for all; the latest evidence from J-PAL affiliated researchers of evaluations of AI innovations across health, governance, labor, climate, agriculture and education; the most promising pathways through which AI can impact social good; and some of the most exciting use cases from innovative organizations.
Impact evaluations, if they're done well, are not just telling us whether a particular AI application worked or not. It is telling us why something worked. How did technology intersect with our lives? How did they intersect with human behavior? How did it change trust in technology? How did it change the ecosystem?
— Iqbal Dhaliwal, Global Executive Director, J-PAL
From Algorithms to Outcomes: Building AI that Works for People
This session set the tone for the day’s discussion, diving deep on aligning research agendas with government priorities so AI adoption and scale-up decisions can be driven by rigorous evidence. Speakers (from left to right): Iqbal Dhaliwal, S. Krishnan, and Michael Kremer
AI in Health: Saving Lives at Scale
Participants learned about the findings from a randomized evaluation in Tamil Nadu testing the impact of a low-cost, AI-based tool to identify people at high risk of heart disease. The subsequent panel discussion unpacked what it takes to make AI innovations in health work in practice and deliver measurable gains. Speakers (from left to right): Rob Sherman, Ziad Obermeyer, Aimee Barnes, Shahed Alam, and Zameer Brey
AI in Governance: Revolutionizing Government Efficiency
This session explored pathways to leverage evidence for effective and equitable delivery of government programs. Participants learned about the impact of using machine learning to improve targeted public service delivery in Togo. This was followed by a discussion on the role of evidence in designing AI tools that strengthen state capacity and broaden citizen participation. Speakers (from left to right): Dean Karlan, Robyn Scott, Kapil Viswanathan, Mohammed Y. Safirulla, and Utkarsh Saxena
AI in Work: Humans, AI, or Both?
This session spotlighted insights from an evaluation of generative AI for teacher hiring in Ghana. It was followed by a panel discussion that examined AI’s profound effects on the job market and the role of scientific evidence in building AI tools that enhance skills, productivity, and worker wellbeing.
Speakers (from left to right): David Yanagizawa-Drott, Becky Faith, Elizabeth Kelly, Shankar Maruwada, and Murugan Vasudevan
AI Vishwaguru: Catalyzing Learning and Impact
J-PAL launched the AI Evidence Playbook at this session, a practical reference for policymakers, practitioners, and donors investing in or developing AI-enabled programs. The launch was followed by a fireside chat on unlocking AI’s full potential for social development through rigorous evaluations and on the lessons India holds for the world on responsible AI deployment. Speakers (from left to right): Audrey Lorvo, Maggie Johnson, Abhishek Singh, Iqbal Dhaliwal, and Sam Carter
AI in Climate and Agriculture: Building Resilient Communities
Participants got an early look at ongoing research studying the impact of an AI tool on reducing flood risk for vulnerable households in Bihar. The panelists then explored the significance and potential of evidence-informed AI applications to drive effective climate action. Speakers (from left to right): Veena Srinivasan, Maulik Jagnani, Fatema AlMulla, Niriksha Shetty, and George Richards
AI in Education: Customizing Learning for All
This session featured insights from a randomized evaluation of an AI-powered essay grader on students’ writing, developed by UNESCO prize winner and education startup Letrus. The panel discussion underscored the importance of scientific evidence in designing and delivering AI tools to help children learn better.
Speakers (from left to right): Sam Carter, Thiago Rached, Brigitte Hoyer Gosselink, Namya Mahajan, Margeurite Clarke, and Rukmini Banerji
Evaluating AI for Impact: From Idea to Execution
The final session brought together key insights from the previous segments, making a compelling case for why impact and evidence should be the foundation for AI for social good innovations. Speaker: Shobhini Mukerji
Launch of the Evidence for AI in Health (EVAH) Initiative
The Summit also marked the launch of the Evidence for AI in Health (EVAH) initiative.
EVAH is a major global effort to generate rigorous evidence to guide the responsible use of artificial intelligence in health systems across low- and middle-income countries (LMICs) in sub-Saharan Africa, South Asia, and Southeast Asia.
EVAH is supported by the Wellcome Trust, the Gates Foundation, and the Novo Nordisk Foundation, and delivered in partnership with J-PAL and the African Population and Health Research Center (APHRC).
Pictured, left to right: Emily Muller (Wellcome Trust), Charlotte Watts (Wellcome Trust), Anthony Francis Mveyange (APHRC), Iqbal Dhaliwal (J-PAL), Hannah Munyao (APHRC), Aimee Barnes (J-PAL), Miranda Barasa (APHRC), and Agnes Kiragga (APHRC).
The Summit Goals House: Keeping the conversation going
From left to right (back): Iqbal Dhaliwal (J-PAL at MIT), Kanishka Narayan (UK Government), and George Richards (Community Jameel). From left to right (front): Owen Larter (Google DeepMind) and Dame Wendy Hall (UN AI advisory body; University of Southampton)
From left to right: Dayoung Lee (Dalberg), Shobhini Mukerji (J-PAL South Asia), and Brigitte Hoyer Gosselink (Google)
J-PAL co-hosted a dinner with Community Jameel and the UK’s Foreign, Commonwealth & Development Office (FCDO) at the India AI Summit Goals House, a venue that featured high-level events running parallel to the India AI Impact Summit 2026. The convening was an opportunity to showcase India and UK’s leadership in redefining the dialogue on global challenges and on AI for social impact. A second dinner, hosted with Google.org as part of Goals House, brought together senior leaders from policy, social impact, technology, and philanthropy to reflect on responsible AI deployment, unlocking new forms of impact, and tackling some of today’s most urgent challenges.
