Artificial intelligence is not only changing the way we write, code, or learn. It is also forcing governments and societies to answer much bigger questions: how are its benefits distributed? What happens to jobs? How is democracy protected? And, above all, who decides the rules?
OpenAI has just announced funding for 14 independent projects seeking to turn those questions into concrete proposals. The initiative provides $1 million in funding and up to $1 million in API credits. The projects will run for six months, and their results will be released in 2027.
Policy must also adapt to AI
OpenAI argues that broad access to technology is only one part of the equation. Although ChatGPT is already used by 1 billion people around the world, that does not guarantee that opportunities will be distributed fairly.
The other part depends on public decisions: how work is organized, how education is funded, how data is protected, and how healthcare, research, and government systems are prepared.
The central idea is simple: technological transformation requires a political transformation of a similar scale.
The company also acknowledges that these decisions should not be left solely to technology companies. In a democracy, public institutions, researchers, social organizations, and citizens should all take part.
That is why the funding is aimed at independent organizations capable of testing ideas, questioning assumptions, and building models that governments can adapt to their own realities.
14 projects to test new solutions
More than 400 people and organizations responded to the call. The selected projects will be developed in the United States, the European Union, Brazil, Singapore, and South Korea, and are divided into two broad areas: expanding economic opportunities and strengthening social resilience.
Work, income, and economic opportunities
-
American Enterprise Institute and Urban Institute: will create low-, medium-, and high-impact scenarios for employment and skills, along with possible responses for each level of disruption.
-
Centre for European Policy Studies: will study how AI-driven productivity gains are distributed across Europe among workers, companies, sectors, and regions. It will also propose new social protection mechanisms.
-
European Centre for International Political Economy: will analyze how people could participate in the ownership of AI’s productive assets through models such as social funds, employee ownership, or citizen investment. The project explores the idea of a possible “right to AI.”
-
Abundance Institute: will investigate how U.S. states can expand their energy generation and transmission capacity to respond to the growth of data centers, while seeking measurable local benefits.
-
Progressive Policy Institute: will design a benefits system linked to the individual rather than exclusively to employment. It would include retirement, healthcare, education, training, leave, and disability support, as well as an insurance model for people whose occupations are transformed.
-
Tax Foundation: will study how AI adoption could change income from labor, business profits, and capital. The goal is to compare different tax options without slowing innovation or weakening fiscal sustainability.
-
Windfall Trust: will expand a network of working groups in the United States, the United Kingdom, Canada, the European Union, and Latin America to analyze employment, welfare, taxes, and competitiveness under different AI progress scenarios.
Science, healthcare, and institutional resilience
-
Instituto de Matemática Pura e Aplicada: will analyze what research institutions need to turn access to AI into real scientific advances, including training, infrastructure, governance, and computing capacity.
-
Hospital das Clínicas of the University of São Paulo: will evaluate a prototype AI-enabled clinical infrastructure for organizing medical information in Brazil’s public healthcare system. It will use synthetic or properly anonymized data and maintain oversight by healthcare professionals.
-
Institute for Security and Technology: will develop a framework for identifying and coordinating responses to AI systems capable of recursively improving their own capabilities. The proposal will include indicators, incident categories, and escalation pathways.
-
Nuclear Threat Initiative: will study how countries can safely share information about capabilities and risks related to the combination of AI and biotechnology, known as
AIxBiorisk. -
Council on Strategic Risks: will train national security professionals in the security and governance of advanced AI systems, with an emphasis on maintaining meaningful human control.
-
Nanyang Technological University: will develop AI agents that simulate how households might respond to public transfers or industrial policies, using behavioral profiles that protect people’s privacy.
-
Yonsei University: will test transparent, human-supervised methods for evaluating the quality of legislative oversight in South Korea’s National Assembly. It will examine the use of evidence, polarization, and the quality of parliamentary oversight.
From broad ideas to concrete tests
The announcement follows up on a proposal published by OpenAI in April 2026 under the title Industrial Policy for the Intelligence Age. In that document, the company argued that governments should use a range of tools, from research and workforce training to regulation and market design.
But there is an important difference between publishing a recommendation and demonstrating that it can work. These projects will try to calculate costs, measure outcomes, create prototypes, and identify risks before certain policies are applied at scale.
For anyone observing AI from outside the technology world, this has a practical meaning. Decisions about artificial intelligence will not determine only which applications we can use. They will also influence our ability to change careers, access healthcare services, participate in the economy, or hold institutions accountable.
OpenAI is funding this initiative, but its results will be produced by external organizations. That independence will be essential so the projects can validate, correct, or even question the company’s own ideas.
The AI era will not be defined solely by more capable models. It will also depend on whether societies can build rules that turn this progress into shared opportunities, with protection for those facing the biggest changes. These 14 projects are a first attempt to expand that conversation beyond technology labs.
Original source
https://openai.com/index/new-policy-ideas-for-the-intelligence-age