OpenAI awards $1M to 14 projects studying AI's economic and social effects

Why it matters

OpenAI is funding policy proposals in areas that will shape its own operating environment, from AI taxation and workforce disruption to data-center energy and model-risk coordination.

OpenAI awarded $1 million to 14 independent projects examining how governments and institutions should respond to AI's effects on employment, public benefits, science, safety and democratic oversight.

The awards, disclosed in an August 17th post from OpenAI Newsroom, also include up to $1 million in API credits across the projects. The recipients will work for six months and report results in 2027, according to OpenAI.

More than 400 people and organizations submitted proposals. That left fewer than 3.5% of applicants in the final group. OpenAI did not break down the cash or API credits by recipient.

The grants turn a set of policy proposals OpenAI published earlier this year into a collection of outside research programs, prototypes and policy frameworks. The work covers the United States, the European Union, Brazil, Singapore and South Korea.

Jobs, benefits and the cost of AI infrastructure

Nine projects focus on economic opportunity. The American Enterprise Institute and Urban Institute will develop scenarios for low, moderate and high disruption to the American workforce, then connect signs of labor-market disruption with possible policy responses.

The Progressive Policy Institute will design a person-based benefits system spanning retirement, health, leave, education, training and disability. Its project includes a prototype for "livelihood insurance" that could be triggered by broad changes within an occupation.

The Tax Foundation will examine how AI adoption could shift government revenue among labor income, corporate profits, capital gains and other sources. The Centre for European Policy Studies will study how productivity gains are distributed among workers, companies, industries and regions, while comparing the European Union with the United States.

OpenAI also selected the Abundance Institute to study how individual US states can expand power generation and transmission in response to data-center demand while producing measurable benefits for local communities. The project will add a policy layer to a public Data Center Atlas, putting infrastructure costs and local economic promises into the same analysis.

Other economic projects will examine wider ownership of productive assets, international coordination on AI economic policy, the requirements for AI-enabled scientific research in Brazil and a clinical information prototype for Brazil's public hospital system. The hospital project will use synthetic or appropriately de-identified data to evaluate clinical workflows and patient services before deployment.

Safety research moves beyond model evaluations

Five projects address what OpenAI calls societal resilience.

The Institute for Security and Technology will develop definitions, warning indicators and incident categories for recursively self-improving AI systems. Its planned framework is intended to connect technical signs of uncontrolled self-modification with coordinated responses by laboratories and governments.

The Nuclear Threat Initiative will examine how countries could securely share information about risks created by the combination of AI and biotechnology. The Council on Strategic Risks will build a frontier AI safety program for national-security professionals.

Nanyang Technological University will develop privacy-preserving AI agents that use financial transaction data to simulate how households might respond to government transfers and industrial policies. Yonsei University will test whether transparent, human-validated AI methods can measure the quality of oversight in South Korea's National Assembly.

OpenAI funds the policy machinery around its models

The program follows OpenAI's Industrial Policy for the Intelligence Age, which proposed research funding, workforce programs, market-shaping measures and targeted regulation as tools for managing increasingly capable AI. OpenAI initially offered focused grants of up to $100,000 as part of the call for proposals.

The resulting portfolio gives OpenAI a way to fund policy development outside its own walls while governments are still deciding how to treat labor displacement, concentrated ownership, infrastructure demand and advanced-model risks. Several recipients are established policy organizations that already speak directly to lawmakers, including the American Enterprise Institute, Urban Institute, Tax Foundation and Nuclear Threat Initiative.

That makes the grants strategically useful despite their limited size. OpenAI is underwriting concrete proposals in the same policy debates that will determine how its technology is taxed, regulated and integrated into public institutions. The selected organizations remain responsible for producing the work, but OpenAI chose the questions and recipients from a pool of more than 400 submissions.

The $1 million cash pool is also a narrow deployment compared with the OpenAI Foundation's larger ambitions. In March, the Foundation said it expected to invest at least $1 billion over the following year across life sciences, jobs and economic impact, AI resilience and community programs.

These 14 projects provide an early view of where that agenda is heading: employment scenarios tied to policy triggers, benefits that follow workers across jobs, new approaches to taxing AI-generated gains, infrastructure planning for data-center power demand and institutional systems for responding to technical risks. The first test will arrive when the projects publish their findings in 2027.