In December, President Trump issued an executive order limiting states’ efforts to regulate artificial intelligence (AI) in favor of a national policy framework to govern AI’s effects on work and life in the US. The order came after 38 states adopted about 100 AI-related measures in 2025.

To date, however, not a single federal law regulating AI has been enacted.

This policy void stands in sharp contrast to American workers' concerns (PDF): About half of workers younger than 34 think AI will affect their job prospects negatively. Among all workers, roughly two-thirds want the government to act on AI to prevent potential job loss, while half believe AI could increase income inequality and polarization.

Meanwhile, AI companies are training their models on the knowledge of billions of human beings.

To help address potential wage and job gaps, policymakers could consider creating universal dividends paid via royalties from AI companies. Universal AI dividends could both help stabilize workers during a potential labor market disruption while recognizing human knowledge as the capital AI companies utilize.

How AI could displace workers

Many tech industry leaders have said that AI will create extraordinary wealth while deeply disrupting the labor market. Some, such as Bill Gates, Sam Altman, Elon Musk, and Mark Zuckerberg, have said some sort of guaranteed income will be needed to support workers during this transition.

Economists widely debate AI’s future effects on the labor market—a debate which current labor market trend data and new forecasting models that test effects of human-AI interaction are currently insufficient to settle.

Some economists think that though AI could disrupt jobs for low-wage and early career workers (ages 22–25), this problem may be offset by more demand for more-skilled workers  (PDF). To date, young workers in the most AI-exposed occupations have experienced a 16 percent decline in employment relative to less exposed occupations.

Other analyses estimate that Al could already replace 12 percent of all workers in the US and could automate tasks that make up 30 percent of US work hours by 2030. Experts from a wide range of disciplines expect the transition to AI to be especially bumpy for lower-wage workers, which could exacerbate skills gaps and income inequality in the US.

If large numbers of workers lose their jobs to AI, they could be retrained for the new economy, but the US has a poor to mixed track record of upskilling displaced workers. Some evidence suggests technological advances are already significantly responsible for reducing the number of quality jobs (PDF) available to US workers.

Treating human knowledge as a public resource

A universal AI dividend could support workers’ transitions in a potentially disrupted labor market and provide shared productivity gains for all citizens.

To turn this concept into policy, policymakers could classify human knowledge itself as the capital AI utilizes and require AI companies to pay royalties into a trust that issues a universal dividend to human beings.

There is precedence for thinking of human knowledge as a public resource.

Albert Einstein once advised students: “Never regard your study as a duty, but as the enviable opportunity to learn to know the liberating influence of beauty in the realm of the spirit for your own personal joy and to the profit of the community to which your later work belongs.” Sam Altman echoed Einstein when he explained that AI systems are trained on the “collective experience, knowledge, [and] learnings of humanity.”

Pete Buttigieg recently spoke of AI dividends, saying they could give “everybody a share in the overall value that is being created by technologies, which, again, rest on technologies that the taxpayer paid for in the first place back in the ’60s,” referencing the federal government’s role in developing the internet. “So why shouldn’t we all get a share?”

Recent reports from the US Copyright Office bend in a similar direction, outlining why only humans should be able to copyright their work under the fair use of copyrighted materials to train AI systems.

How policymakers could use lessons from the Alaska Permanent Fund to create a universal AI dividend

In the past, policymakers have required industries to compensate the public for the use of a public resource. As the Alaska Permanent Fund (APF) has demonstrated, wealth generated by a shared resource—in its case, oil and gas—can be distributed directly back to citizens universally, fairly, and with bipartisan support.

Mineral rights for land in Alaska are owned by the state, the US government, and the Alaska Native corporations and individual Native landowners under the Alaska Native Claims Settlement Act. Since 1976, 25 percent of revenue from oil and mineral leases on state-owned lands and from federal mineral revenue-sharing payments are paid as royalties and deposited directly into the APF for investment and distribution to citizens of Alaska.

Each Alaskan citizen, including children, receives an annual dividend payment (typically $1,000 to $2,000) based on a five-year average of the fund’s performance. As a result, the APF has reduced the number of Alaskans with incomes below the US poverty threshold by 20 to 40 percent, with even larger effects for historically oppressed populations, seniors, and children. And the APF has done this without substantially affecting labor market participation.

The APF is one of many examples of how guaranteed income can help stabilize workers in the event of widespread job or wage loss. In addition, research on other guaranteed income programs has demonstrated excellent outcomes for children; substantial gains in adult mental health, physical health, and parenting skills; and increased food security and housing stability. 

Even though AI products aren’t generating substantial revenues yet, research suggests they will soon. To prepare for the future of AI, Congress and states should convene task forces and advance research that examines how to design and implement a national AI permanent fund.

These taskforces could consider the following:

• the share of revenue that AI companies should pay in royalties
• a progressive system for dividend distribution that pays out larger amounts to displaced workers and other populations in need
• the share of royalties that should go to other public goods, like worker retraining
• strategies to adjust distribution in the event of a widespread downturn in the AI sector

States, which have long been laboratories for evidence-based policy innovation, can help test various policy and program options for a national AI permanent fund. Organizations that already convene state policymakers could assist by creating a commission to define the proper state-federal balance for AI laws and regulations.

AI companies train their models on the ideas, stories, data, and labor of billions of human beings—including the Einsteins of the world, you, and me. A national AI permanent fund that pays out a dividend to all citizens could be the next logical step for democratic capitalism in a changing world. It would be the policy equivalent of saying: If AI is built on what all of us own, then all of us should benefit.