OpenAI funds 14 independent policy projects on AI opportunity and resilience

OpenAI is funding 14 independent policy projects intended to examine how advanced AI could broaden economic opportunity and strengthen social resilience. The August 17 announcement follows the company’s April paper, Industrial Policy for the Intelligence Age, and turns part of that agenda over to outside organizations that can test, challenge and refine policy ideas beyond OpenAI’s own institutional view.

The company argues that access to AI alone will not determine whether the technology’s benefits are widely shared. OpenAI says ChatGPT is now used by 1 billion people around the world, spanning geographies, income levels, demographics and major adult age groups. But the announcement emphasizes that deployment choices, public institutions and democratic debate will shape whether AI improves opportunity broadly or concentrates gains among a smaller set of firms, workers and regions.

OpenAI says more than 400 people and organizations responded to its call for proposals. The selected projects will receive a total of 1 million dollars in funding and up to 1 million dollars in API credits. They span the United States, the European Union, Brazil, Singapore and South Korea, and the work ranges from research papers and policy models to datasets, prototypes and practical frameworks that can be tested by governments and institutions.

Several projects focus on economic participation. The American Enterprise Institute and Urban Institute will convene a bipartisan commission on AI and the future of the workforce, while the Centre for European Policy Studies will study how AI-driven productivity gains are distributed across workers, firms, sectors and regions in Europe. ECIPE will examine ideas such as a Right to AI and broader ownership of productive assets. The Abundance Institute will look at state-level energy policy for data center demand, and the Tax Foundation will study how AI could affect public revenue.

Other grants move closer to public services and institutional readiness. Brazil’s IMPA will study what research organizations need to turn AI access into scientific progress. Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo will evaluate a clinical information infrastructure prototype for Brazil’s public health system using synthetic or de-identified data. Nanyang Technological University will build privacy-preserving LLM agents for auditable government policy simulation, while Yonsei University will test AI-enabled methods for assessing democratic accountability in South Korea’s National Assembly.

The resilience side of the program shows how policy conversations are widening beyond jobs. The Institute for Security and Technology will work on a framework for measuring and coordinating responses to uncontrolled recursively self-improving systems. NTI bio will explore cross-border information sharing for AI and biology risks. The Council on Strategic Risks will develop a frontier AI safety module and lab immersion program for national-security professionals. These projects reflect a shift from abstract safety debate toward institutional playbooks and observable signals.

The grants are not a product launch, but they matter because policy experiments often become procurement rules, disclosure practices, public-sector pilots and eventually regulation. OpenAI says the projects will run for six months, with results reported in 2027. The broader signal is that major AI developers know they cannot define the economic and governance agenda alone. For companies building AI tools, the work could influence future expectations around worker transition, energy siting, clinical oversight, public-sector simulation and frontier-risk reporting. For users, it is another reminder that the next phase of AI will be shaped not only by model capability, but by the institutions built around deployment.