A nonprofit R&D lab that develops collective intelligence tools and governance models to steer transformative AI development toward better outcomes through democratic public input.
A nonprofit R&D lab that develops collective intelligence tools and governance models to steer transformative AI development toward better outcomes through democratic public input.
People
Updated 05/18/26Co-Founder & Executive Director
Co-Founder & Technology Advisor
Research Director
Senior Research Fellow
Head of Global Partnerships
Communications and Operations Director
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The Collective Intelligence Project (CIP) is a nonprofit R&D organization founded in August 2022 by Divya Siddarth and Saffron Huang, with Jasmine Wang as a third co-founder. The organization publicly launched with a whitepaper in February 2023 that outlined its vision for new governance models for transformative technology. CIP is headquartered in San Francisco, having opened its first office in Japantown in 2024. CIP's central thesis is that existing governance approaches for transformative technologies force trade-offs between progress, safety, and democratic participation, what they call the Transformative Technology Trilemma. CIP works to develop a fourth path through collective intelligence capabilities that integrate all three priorities. The organization operates several flagship programs. Global Dialogues is a platform for structured deliberation that brings diverse global voices into AI development decisions. In 2025, CIP conducted seven rounds of Global Dialogues engaging over 10,000 participants across more than 70 countries. Weval is an open platform enabling communities to create qualitative benchmarks and evaluations for AI systems. Community Models allows communities to create and refine AI models based on collectively defined constitutions. Samiksha, launched in partnership with Karya and Microsoft Research, generates evaluations of AI systems across Indian languages and domains. CIP has partnered with major AI labs including OpenAI, Anthropic, and Mistral AI, as well as governments in Taiwan, Bhutan, and Montenegro on national AI strategies. The organization ran alignment assemblies in 2023-2024 alongside committed partners to incorporate public voice into AI decision-making. CIP's co-founders Saffron Huang and Divya Siddarth were named to TIME's 100 Most Influential People in AI in 2024. The organization is a 501(c)(3) nonprofit (EIN 92-0327339) that operates exclusively through philanthropic charitable donations. CIP's funders include the Robert Wood Johnson Foundation, Google.org, Ford Foundation, Omidyar Network, One Project, Future of Life Institute, Amaranth Foundation, and the Survival and Flourishing Fund. Before achieving its own 501(c)(3) status in 2024, CIP was fiscally sponsored by the RadicalxChange Foundation. The team includes notable figures such as Audrey Tang, Taiwan's first Digital Minister (2016-2024), who joined as Senior Research Fellow, and Zarinah Agnew, a neuroscientist who serves as Research Director. CIP is featured in publications including TIME, Wired, Nature, BBC, The New York Times, and Science.
Theory of Change
Updated 05/18/26CIP identifies a Transformative Technology Trilemma in which existing governance approaches force trade-offs between technological progress, safety, and democratic participation. Their theory of change is that by developing and deploying collective intelligence tools -- scalable processes for surfacing, aggregating, and acting on diverse public values -- they can create a fourth path that integrates all three priorities. Concretely, CIP builds platforms for large-scale public deliberation (Global Dialogues), community-driven AI evaluation (Weval), and participatory model alignment (Community Models), then partners with AI labs and governments to embed the resulting insights into actual development and policy decisions. This creates a feedback loop where public input directly influences how AI systems are built, evaluated, and governed, reducing the risk that transformative AI is developed without adequate democratic accountability or diverse safety considerations.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26A research and policy roadmap that lays out concrete steps stakeholders can take in the near term to build a more democratic AI ecosystem that is adaptive, accountable, and aligned with public values.
A series of democratic deliberation processes that bring diverse publics into questions like AI release, risk, and governance, translating their preferences into guidance for labs, governments, and civil-society partners.
A six‑month, fully remote research fellowship for three early‑career scholars to use CIP’s proprietary datasets to produce original work on AI, public values, and democratic governance.
Discussion
No comments yet. Be the first to share your thoughts.