Bernie Sanders’ recent proposal that half of artificial intelligence should be owned by the public is unlikely to become policy anytime soon, but it reflects a broader debate that is gaining momentum among economists, technology researchers, and policymakers: How will Americans benefit if AI creates trillions of dollars of new economic value? There is no shortage of ideas, but most of them are untested. However, the risks and growing public backlash against AI have made this issue an important one.
Although AI wealth has accumulated rapidly in the stock market, the amount that many Americans can benefit from that growth is still limited. Recent research shows that a majority of American workers now want to hold companies more accountable through AI sovereign wealth funds. There are unconfirmed reports that OpenAI is considering offering a 5% stake to the government ahead of its highly anticipated IPO. Meanwhile, Jeff Bezos recently told CNBC that the best policy idea to level the economic playing field is simply to eliminate federal income taxes on the bottom half of American earners.
Recent survey data shows that this question is embedded in rapidly changing public sentiment towards AI. An Emerson College poll released this week found that only 27% of Americans support building data centers in or near their communities, while 63% oppose it. National sentiment has deteriorated significantly in less than a year. A similar poll conducted in December 2025 found that 33% supported such a development, while only 42% opposed it. Many Americans feel as if they have nothing to gain and everything to lose from AI.
“I look at the data center proposals and see no progress,” Northeast Ohio resident Will Hollingsworth said during an April public comment session on a proposed 257-acre data center campus in Portage County. “I see a gamble where big tech companies get the money and Portage County pays for it.”
“We’re being asked to sacrifice the lifeblood of our city so a trillion-dollar company can save a few pennies on their margins,” Hollingsworth said in a comment that went viral. “We’re being asked to drain reservoirs so chatbots can write poems and such, so our sheriff can generate a picture of himself standing next to a Bigfoot.”
Several proposals exist among economists, technology industry researchers, and public policy experts to address Hollingsworth’s sentiment that the potential outcomes of AI development are heavily skewed in favor of corporations. These include partial public ownership models in AI and other shared equity mechanisms.
Computer scientist Jaron Lanier, currently Chief Technology Officer and Principal Unified Scientist at Microsoft Research, advocates for a model, also known as “data dignity,” in which people receive compensation for the information and contributions that help build AI systems.
“I spent some time with Sen. Sanders when he visited the AI community at Stanford University,” Lanier said. “Whether[his proposal]is a good idea depends on the nature of the government responsible for distributing benefits to the people,” he said. If the government were to become “just another AI company,” he would prefer what he calls a more “decentralized economic model.”
“With the right data and oversight, there could be enough real money to make a significant impact on people’s lives,” Lanier said.
“But if the future is going to look like a standard Silicon Valley, where people’s contributions are anonymized, people are dismissed as if AI did all the work, and people are fictionalized as useless, then it’s better to have some support through government structures that have participatory and democratic elements,” Lanier said.
Pay people directly for AI training data
The challenge is understanding how such systems work. AI models are trained on vast amounts of information from millions or billions of sources. Determining which individual contributions created value and how much they should be paid for their contributions may run into the same criticisms as persistent efforts to compensate people for their search histories. Although the total amount is huge, the economic value of personal data is low.
Raul Castro Fernández, an assistant professor of computer science at the University of Chicago, recently wrote about how to fairly compensate the public for AI, refuting the argument that it is impossible to accurately track (and compensate) the vast amount of data points collected by AI models from human contributors. “The most powerful version of profit sharing is not a tax, but a reward system tied to the human contributions that make the AI system valuable in the first place,” Fernandez said.
“They (AI companies) are already estimating how important the data is through the laws of scaling,” Fernandez said. “A plausible mechanism would be closer in spirit to a collective control system similar to music royalties than to calculating the exact value of individual ‘tokens,'” he said. “AI companies would pay a portion of their model’s profits into a pool, the total share would be determined based on evidence about how dependent the model’s performance is on data, and payments would be distributed among creators, publishers, platforms, and other intermediaries according to an audited measure of data contribution.”
But researchers Nicholas Vincent and Brent Hecht from Simon Fraser University and Northwestern University, respectively, warn against this approach. In a 2023 study that investigated whether it is possible to appropriately reward individuals for their contributions to AI systems, Vincent and Hecht argue that assigning a rating to each individual’s data can be highly subjective and counterintuitive.
“Seemingly trivial design choices can significantly alter the distribution of data values, which is a serious concern for human-AI systems that seek to incorporate such values for payments and other purposes,” they said. “If a technology relies on the collective contributions of millions or even billions of people, why spend time and energy estimating (potentially costly) data values when we already know that the individual values are very small?” they concluded.
Creating a new, powerful trade union for the 21st century
However, direct payments may not be the only way to achieve a fairer data procurement process. Matt Prewitt, president of the RadicalxChange Foundation and one of the two authors of the policy paper, advocates for creating a new kind of legal right that gives people the power to shape how AI works. This is the 21st century version of trade unionism, where “people cannot waive these rights at an individual level. Instead, people must join an association to exercise these rights.”
This will create a new type of regulated entity that will have “a very important seat at the table with AI companies and the power to capture equity, compensation, governance and power,” Prewitt said.
Economist and engineer Glenn Weil, principal scientist at Microsoft and founder of RadicalxChange, argues that goals should not necessarily be in the public domain of governments or businesses. According to RadicalxChange, efforts to “fragment and fragment ownership (i.e., give some of the traditional ownership rights to more or different people) or consolidate ownership (i.e., place it in the hands of national representatives such as states)” have had some effect, but are best viewed as “mere band-aids.”
“Ownership fragmentation only ‘spreads around’ the same old exploitative incentives of traditional property rights, whereas property consolidation ‘puts all your eggs in one basket’ and increases the risk of institutional capture and illegal representation,” RadicalxChange staff wrote in a policy paper advocating for a new model centered on collective ownership.
New corporate tax, reduced working hours
Some say mechanisms already exist for policymakers to build a fairer AI economy without relying on untested ideas. Those include stronger corporate taxes, antitrust enforcement and labor protections, said Dean Baker, an economist and co-founder of the Center for Economic Policy Research.
Baker said he remains unconvinced that AI will take over a large amount of human labor, but added that he would still rely on “old cures.”
Those relief measures could include “a higher rate of enforceable corporate income taxes for all businesses,” Baker said. But payment methods could be new, he added. “The best way to do this is to require companies to surrender non-voting stock equal to the target tax rate (e.g. 25% of the stock if the tax rate is 25%),” he said.
Additionally, Baker says that strict enforcement of antitrust laws would provide a plausible route to more equitably economically sharing the benefits of AI. Baker offered an analogy about cheap Chinese products replacing blue-collar labor. “We’re distorting a large part of the blue-collar workforce by bringing in Chinese products. We shouldn’t be doing protectionism to keep Elon Musk and Mark Zuckerberg incredibly rich.”
Avoiding repeating past policy failures, such as the rise of social media and a lax attitude toward the early global outsourcing era, has captured the attention of some of America’s most senior policymakers, who are betting that AI, especially the employment aspect, will grow as an electoral and social issue in the coming years.
While the idea of a universal basic income, or what Elon Musk calls a “universal high income” program, to combat mass unemployment has been discussed for years, simpler labor market mechanisms for distributing future economic efficiencies created by AI already have global precedent.
The answer is not to work less, but to work less.
“We instituted a 40-hour work week 90 years ago and it hasn’t changed since then,” Baker said. “Other countries are shrinking workweeks and working hours per year. If AI delivers the productivity gains promised, let’s lower the threshold to 32 hours or even lower. We could even double the overtime premium to 100% instead of 50%,” Baker said.
