Bernie Sanders’ recent proposal that society should have an artificial intelligence component is unlikely to become policy anytime soon, but it does reflect a broader debate that is raging among economists, technology researchers, and policymakers: How can Americans benefit if AI creates billions of dollars in new economic value? There is no shortage of ideas, most of them untested. But the risks, and the rise of public opposition to AI, make this question even more important.
AI fortunes have quickly amassed in the stock market, but many Americans are still limited in how much they benefit from that growth. Recent survey work shows that a majority of American workers now want to hold companies accountable for the AI economy. There have been unconfirmed reports that ahead of the much-anticipated IPO, OpenAI has discussed giving the government a 5% equity stake. Meanwhile, Jeff Bezos recently told CNBC that the best policy idea to level the economic playing field is to eliminate the income tax for the bottom half of US earners.
Recent survey data shows that this question is embedded in a rapid shift in public sentiment about AI. An Emerson College poll released this week found that only 27% of Americans support data centers being built in or near their community, while 63% oppose. Public sentiment has run high in less than a year. A similar survey conducted in December 2025 found that although 33% said they would support this, only 42% expressed their opposition. Many Americans feel like they have nothing to gain and everything to lose from AI.
“When I see a data center proposal, I don’t see progress,” said Will Hollingsworth, a Northeast Ohio resident, speaking at an April public comment hearing about the 257-acre data center in Portage County. “I see a gamble where the big tech companies strike gold while Portage County foots the bill.”
“We’re being asked to donate the blood of our city so that a multi-billion dollar company can save half a cent on its bottom line,” Hollingsworth said in the opening remarks. “We are asked to release our water like that [that] a chatbot can write a poem or something like that [that] Our sheriff can produce a picture of himself standing next to Bigfoot.”
Among economists, technology industry researchers, and public policy experts, there are many proposals that address Hollingsworth’s feeling that the potential outcomes from AI development are significantly skewed in favor of corporations. These include models of partial public ownership in AI and other forms of equity sharing.
Computer scientist Jaron Lanier, who currently holds the Office of Chief Technology Officer and Premier Associate Scientist at Microsoft Research, spoke about a model sometimes called “data dignity,” where people are compensated for information and contributions that help create AI programs.
“I spent some time with Sen. Sanders when he visited the AI community at Stanford,” Lanier said. “Or [his proposal] it would be a good idea depending on the state of the government that will be responsible for delivering the benefits to the people.” If the government could become “just another AI company,” he prefers what he calls a “distributed economic model.”
“Good data and monitoring,” Lanier said, can lead to enough real money to have a big impact on people’s lives.
“But if the future is going to be a typical Silicon Valley, where people are going to be mythologized as useless because their contributions have been exposed and dismissed to make it look like AI has done all the work, it’s better that some kind of support comes from a participatory/democratic government structure,” Lanier said.
Paying people directly for AI training data
The challenge is figuring out how this system will work. AI models are trained on large amounts of information from millions or billions of sources. Determining which individual contributions have created value, and how much should be paid, may run into the same criticism as radical attempts to compensate people for search histories – while the numbers are large when aggregated, the economic value of any one individual’s data is low.
Raul Castro Fernandez, an assistant professor of computer science at the University of Chicago who recently wrote about how to properly compensate society for AI, refutes the argument that it is not possible to accurately track (and compensate) the large amounts of data points collected by AI models from human contributors. “The most powerful version of profit sharing is not a tax but a compensation system that complements the human contributions that make AI systems valuable in the first place,” Fernandez said.
“See [the AI companies] you just estimate how much data is important with the rules of scaling,” said Fernandez. “A reasonable way would look less like calculating the exact amount of each ‘token’ and more like a collective management system, similar in spirit to music awards: AI companies will pay part of the model profit in the pool, the aggregate amount through payment shares would depend on how much the aggregate amount would depend on, and distribution of data. to all creators, publishers, platforms, or other intermediaries according to the tested measures to contribute data,” he said.
But Nicholas Vincent and Brent Hecht, researchers at Simon Fraser University and Northwestern University, respectively, caution against this approach. In a 2023 study examining whether it is possible to adequately compensate individuals for their contributions to an AI system, Vincent and Hecht argue that attaching ratings to individual data may be too subjective and potentially counterproductive.
“Seemingly small design choices can significantly change the distribution of data values, a major concern for any human AI program that wants to aggregate such values for payments or other purposes,” they said. “If technology depends on the collective investment of millions or billions of people, we already know that each amount will be very small, so why bother spending time and energy working on it [potentially costly] quantifying the amount of data?” they concluded.
Creating powerful new unions for the 21st century
Direct payment may not be the only way to achieve a fair data access process, however. Matt Prewitt, president of the RadicalxChange Foundation and one of the two authors of that policy paper, is advocating for the creation of a new category of legal rights that give people the power to shape how AI works, a 21st century version of unions where “people can’t sign away these rights on an individual basis. Instead, people have to join together in unions to exercise these rights.”
This will create a new class of regulated entities that “have a very serious seat at the table with AI companies, and have the power to acquire shares, income, governance, and power,” Prewitt said.
Economist and expert Glen Weyl, principal researcher at Microsoft and founder of RadicalxChange, argues that the goal should not necessarily be government or public ownership of corporations. According to RadicalxChange, efforts to “segregate and divide ownership (that is, to give many or different people a share of common ownership); or to consolidate ownership (that is, to put it in the hands of some public agent, such as the state),” can do good but are best seen as “only musical instruments.”
“Diversifying ownership simply ‘spreads’ the same old motivations of common ownership, while consolidating ownership ‘puts all eggs in one basket,’ strengthening the risk of institutional capture and illegal representation,” RadicalxChange staff wrote in a policy piece opposing new models focused on common ownership.
New business taxes, shorter working hours
Others say that ways already exist for policymakers to build a balanced AI economy without resorting to untested ideas. According to Dean Baker, economist and founder of the Center for Economic and Policy Research, these include stronger corporate taxes, antitrust enforcement, and labor protections.
While Baker said he is not yet convinced that there will be mass displacement by AI, he added that he would still go back to “the old remedies.”
These remedies would include “a higher effective corporate income tax, for all companies,” Baker said. But he added that the payment method could be new. “The best way to do this is for companies to convert non-voting shares equal to the target tax rate (eg 25% shares at a 25% tax rate”), he said.
Additionally, Baker argues that if applied strictly, antitrust law provides a sound channel for the economically fair distribution of AI benefits. Baker gave the analogy of cheap Chinese products displacing blue-collar workers. “We’re allowing Chinese-made goods to absorb large parts of the workforce. We shouldn’t have protectionism to keep Elon Musk and Mark Zuckerberg ridiculously rich,” he said.
Not repeating the policy mistakes of the past, including a dismissive attitude toward the rise of social media and the previous era of global outsourcing, is high on the radar of some of the most senior policy makers in the US, who are betting that AI, and the jobs aspect in particular, will grow as an electoral and social issue in the coming years.
Although the idea of a universal basic income – or a system of “global high income” as Elon Musk calls it – to fight mass unemployment has been tied up for years, there is a simple labor market way to spread the economic efficiency of the future created by AI that already has a global precedent.
The answer is not work, but less work.
“We set the 40-hour work week 90 years ago and it hasn’t changed since,” Baker said. “Some countries have shortened the work week, the work year. If AI will give us the promised boom in productivity, let’s lower the limit to 32 hours, or maybe even lower. We can also double the overtime premium to 100% rather than 50%,” said Baker.
