Divya Siddarth

Divya Siddarth is developing practical ways for public judgment to enter the design and evaluation of AI. Her work turns deliberation into datasets, evaluation criteria and institutional processes, seeking technological governance that reflects lived experience rather than expert preference alone.

Divya Siddarth approaches AI governance as an institutional-design problem: if systems will mediate public life, the public needs credible ways to shape what those systems are asked to optimise. As Co-Founder and Executive Director of the Collective Intelligence Project, she has helped build deliberative processes that translate diverse views into research and evaluation artefacts. The organisation’s 2025 Global Dialogues Index drew on seven rounds involving more than 6,000 participants across 70 countries, examining how people use AI, where trust accumulates and where expectations diverge. Its Weval platform takes a complementary route, enabling communities and domain specialists to construct qualitative evaluations around situated concerns that standard benchmarks often miss. In India, the organisation’s Samiksha programme reported more than 25,000 queries in 11 languages and over 100,000 manual evaluations across health, agriculture, education and law. Siddarth’s contribution is best understood as institution-building rather than personal authorship of every output. She has helped give public participation a technical pathway—from deliberation, to criteria, to tests that developers and public authorities can use.