By: Jiayi Jessica Ye, Yanmei Lin

Every prompt submitted to ChatGPT, every Wikipedia edit, every online article, and every piece of publicly accessible knowledge contributes to the digital commons on which modern artificial intelligence (AI) is built. Yet the extraordinary economic value generated from these collective contributions increasingly flows to a handful of technology companies rather than to the billions of people whose knowledge makes AI possible. This raises a deceptively simple legal question: if AI derives its value from a shared intellectual common, who should benefit from its commercialization? The debate is often framed around copyright, data licensing, or AI regulation. But more than two centuries ago, Thomas Paine confronted a remarkably similar problem. Although writing about land rather than data, Paine developed a theory of property that speaks directly to today’s AI economy.

In 1797, as Europe was overwhelmed by enclosure and revolution, Bishop Richard Watson contended that the division between rich and poor was “appointed by God”.[1] However, Thomas Paine argued that God only made male and female, and poverty is a product of “civilized life” under a specific property regime.[2]

While John Locke argued that labor creates title to land, Paine insisted that “It is impossible to separate the improvement made by cultivation from the earth itself”.[3] The landowner therefore owes society a “ground rent” for enclosing what was once common inheritance. Thus, the agrarian justice that Paine is proposing is a rights-based theory of compensation.

This blog argues that Paine’s theory provides an overlooked framework for governing artificial intelligence. If landowners owed society ground-rent for enclosing common land, AI developers owe society what might be called compute-rent, compensation for enclosing and commercializing humanity’s shared intellectual commons. Universal Basic Income (UBI), viewed through this lens, is therefore not merely social welfare but a modern mechanism of restitution.

If ground-rent is the price paid for enclosing common land, compute-rent is the price AI developers should pay for enclosing the digital commons that makes their models. Large Language Models (LLMs) are built upon resources from massive digital commons. GPT-4 was trained on trillions of data drawn from public web and academic archives.[4] As Paine contended, without the natural land, the improvement is impossible. Similarly, without all these human collective data, AI models are just hollow shells. This functional inseparability is what urges for a restitution mechanism. Big firms are currently free riding on the original value of human input, and because computational power is highly concentrated, firms like Microsoft, Google and Meta have become de facto digital landlords.[5] In 2025, the five largest firms, OpenAI, Anthropic, NVIDIA, Adobe and Microsoft, together held roughly 50 percent of the market.[6] These firms have been exercising exclusive control, benefiting from a surplus they didn’t create and having terms of exchange that were never consensually negotiated with the contributors.

Paine’s argument revealed further impacts beyond the loss of property with peasants back then losing the possibility of exchanging labor for subsistence. In the current context, critics like economist Hal Varian and Carl Shapiro argued that information doesn’t deplete when used.[7] When an AI company gets information from a public blog, the original author can still benefit from it. Therefore, there’s no exclusion in the way there was during the Enclosure Acts.[8] However, while data is non-rivalrous, the economic opportunity derived from it is rivalrous. If an LLM can simulate a writer’s style, it occupies the market spots previously held by humans. We see this in empirical numbers. The International Monetary Fund estimates that about 60 percent of jobs in advanced economies are at risk due to AI,[9] and by 2030, as many as 375 million workers across the globe will face job disruption and need to switch occupations.[10]

A pointed objection might be that compute-rent weakens innovation incentives. AI firms invest heavily in infrastructure, treating public data as a compensable common may raise costs and slow deployment. Yet Paine’s framework doesn’t deny returns on improvement. It only insists that improvement don’t erase the commons that made it possible.

Rights-Based Governance of the Digital Commons

Paine’s argument in 1797 was rigorous, but it remained just a theoretical argument. His proposal of the national fund was never enacted since the government in charge were the wealthy landowners who benefited from privatizing the land.[11] The nineteenth-century industrial and colonial expansion that followed allowed Britain to defer the problem by exporting its costs elsewhere. Yet the AI era leaves us with nowhere to hide, absorbing human knowledge instantly, and we cannot simply “export” the economic displacement and job disruption it creates. What it encloses cannot be returned either: once the cognitive labor of billions has been compressed into a model’s parameters, that labor cannot be extracted, restored, or redistributed.[12][13] Each year of policy delay adds more to those weights, and neither of the two governance forms now visible across jurisdictions can resolve this. For market mechanisms, anonymous contributors (like Wikipedia editors or historical writers) cannot band together to negotiate a payout, and AI developers have no incentive to compensate the general public if they can just scrape the data for free.[14] For state-led mechanisms, if governments intervene, they often hoard the money to fund state projects, simply leaving the commons enclosed by a different gatekeeper.[15]

What is required is what Paine envisioned of public trust, holding compute-rent on behalf of those whose contribution made it possible, disbursing it universally by right, bypassing corporations and politicians entirely. This form has a name in the twenty-first century: Universal Basic Income.

UBI as a Pathway to Rights-Based Justice

Universal Basic Income (UBI), an unconditional payment to every citizen by the government, would be the policy form this restitution takes, acting as the compensation for exclusion from a common resource.[16] Currently, most social safety nets are funded through labor taxation. But in this context, if compensation is funded through labor taxation, those displaced by technology are forced to pay for the systems that replace them. In fact, as AI replaces human tasks, the income tax base will naturally shrink; attempting to fund a social safety net by increasing taxes on a disappearing workforce is by itself counterproductive. Instead, a modern national fund should be financed through a data dividend, with taxes on excess profits or licensing fees on public data trained models to act as recognition of every citizen’s legal right.

But why universal? This is because as contributions are socially produced rather than individually traceable, any non-universal mechanism would reintroduce the original problem. Consider two mechanisms we have currently. First are litigation cases like The New York Times v. Microsoft Corp.[17] If such cases succeed, compensation directly flows to official holders (newspapers, image archives, publishers), but not to contributors whose language and knowledge are cultivating the models indirectly. The second is the data licensing agreements. OpenAI and Google pay Reddit about $130 million per year combined to legally access Reddit’s data.[18][19] In that case, the corporate platform takes the entire payout, while the everyday users who actually wrote the content receive $0. Both reintroduce the attribution problem, but a universal dividend would resolve it by tying the dividend to citizenship, making the adjudication of “who counts” unnecessary. By grounding UBI, we renew the social contract for an era where human labor is no longer the primary factor of production.

Legitimacy and the Stability of Property

Failing to renew this social contract, however, may trigger a crisis of institutional legitimacy. On a broader scale, Paine also argues that when most people judge a property system to be unjust, the system becomes fragile. He wrote that throughout Europe the state of civilization was “unjust” in both principle and effect, and that this injustice, once perceived, made those property holders themselves live in the fear of the very system that sustained them.[20] The same logic applies in the age of AI. If billions of people come to realize that their language, creativity, and idea contributions are used to train systems generating trillions in value while they themselves are economically marginalized, they will begin to question the legitimacy of both those particular firms and the property rights protection regime that allows the enclosure. As “AI reshapes labor markets, concentrates wealth and data governance power,” it creates immense inequality leading to the society facing a legitimacy crisis.[21][22] Indeed, scholars argue that because AI creates massive economic rents for information monopolies, failing to redistribute these gains will lead to severe societal friction;[23] without more inclusive institutional frameworks, this rapid technological displacement could bring social unrest.[24][25]

This frames UBI in a way that should appeal even to those who care most about protecting property. As Paine argued, stabilizing property through institutionalized justice is cheaper and more durable than maintaining it through coercion. A data dividend funded by compute-rent is by itself a defense in property rights by ensuring that everyone receives a foundational economic floor sufficient to ensure basic standards of living within the social system, protecting the legitimacy and security of the existing system.

Towards a New Social Contract

If the productivity gains of AI concentrate without restitution, firms may optimize individually while the social conditions would erode gradually. Returning to Paine’s insistence on rights offers a new social contract that treats the digital commons as a shared resource. Through compute-rent, we stop people from being unpaid indirect labor for tech platforms and become legal shareholders in the AI economy. If eighteenth-century landowners owed ground-rent, twenty-first-century AI developers owe compute-rent.

 

[1] Richard Watson, The Wisdom and Goodness of God (1793), https://anglicanhistory.org/wales/watson_wisdom1793.html.

[2] Paul Meany, Thomas Paine’s Solution for Poverty, LIBERTARIANISM.ORG (Aug. 3, 2020), https://www.libertarianism.org/articles/thomas-paines-solution-poverty.

[3] Thomas Paine, Agrarian Justice (1797), https://www.ssa.gov/history/paine4.html.

[4] Dingyuan Luo et al., Evaluating the Performance of GPT-3.5, GPT-4, and GPT-4o in the Chinese National Medical Licensing Examination, 15 SCI. REP. 14119 (2025), https://doi.org/10.1038/s41598-025-98949-2.

[5] Guillaume Belin, AI Transforms Cloud Giants into New Landlords, JOURNAL D’UN PROGRESSISTE (May 25, 2026), https://journaldunprogressiste.fr/en/ai-transforms-cloud-giants-into-new-landlords-en/.

[6] MarketsandMarkets, Top Companies in Generative AI Market—Microsoft (US), Google (US), IBM (US), NVIDIA (US) and OpenAI (US), https://www.marketsandmarkets.com/ResearchInsight/generative-ai-market.asp.

[7] Carl Shapiro & Hal R. Varian, Information Rules: A Strategic Guide to the Network Economy (1999), https://www.kcmit.edu.np/Uploads/information-rulesLarge20210211052224.pdf.

[8] UK Parliament, Enclosing the Land, https://www.parliament.uk/about/living-heritage/transformingsociety/towncountry/landscape/overview/enclosingland/.

[9] Mauro Cazzaniga et al., Gen-AI: Artificial Intelligence and the Future of Work, IMF Staff Discussion Note No. 2024/001 (2024), https://doi.org/10.5089/9798400262548.006.

[10] James Manyika et al., Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation (McKinsey Glob. Inst. 2017), https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages.

[11] Brent Ranalli, Thomas Paine’s Agrarian Justice and John Locke, THE GLOBALIST (July 3, 2015), https://www.theglobalist.com/thomas-paine-agrarian-justice-and-john-locke/.

[12] Pieter Verdegem, Dismantling AI Capitalism: The Commons as an Alternative to the Power Concentration of Big Tech, 39 AI & SOC’Y 727 (2024), https://doi.org/10.1007/s00146-022-01437-8.

[13] Wenjuan Jia & Wenxi Yan, Labor Control in Cognitive Labor and Data Labeling: The Case of AI Company N, 13 J. CHINESE SOC. art. 5 (2026), https://doi.org/10.1186/s40711-025-00253-z.

[14] Jing Hu & Graham Lovelace, The Prisoner’s Dilemma Between AI and Creators, 2ND ORDER THINKERS (Apr. 12, 2025), https://www.2ndorderthinkers.com/p/the-prisoners-dilemma-between-ai.

[15] Janet Vertesi et al., Reckoning with the Political Economy of AI: Avoiding Decoys in Pursuit of Accountability, in Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency 2186 (2026), https://doi.org/10.1145/3805689.3806739.

[16] Bernard Marr, Will AI Make Universal Basic Income Inevitable?, FORBES (Dec. 12, 2024), https://www.forbes.com/sites/bernardmarr/2024/12/12/will-ai-make-universal-basic-income-inevitable/.

[17] David Grossman & Elena De Santis, New York Times v. Microsoft Corp., LOEB & LOEB LLP, https://www.loeb.com/en/insights/publications/2025/04/new-york-times-v-microsoft-corp.

[18] Anna Tong et al., Exclusive: Reddit in AI Content Licensing Deal with Google, REUTERS (Feb. 22, 2024), https://www.reuters.com/technology/reddit-ai-content-licensing-deal-with-google-sources-say-2024-02-22/.

[19] Kendra Barnett, AI Licensing Deals with Google and OpenAI Make Up Nearly 10% of Reddit’s Revenue, ADWEEK (Feb. 13, 2025), https://www.adweek.com/social-marketing/ai-licensing-deals-with-google-and-openai-make-up-10-of-reddits-revenue/.

[20] Brent Ranalli, Thomas Paine’s 1797 Call for a Basic Income: A New Paper Tells the Full Story, RESILIENCE (Apr. 20, 2020), https://www.resilience.org/stories/2020-04-20/thomas-paines-1797-call-for-a-basic-income-a-new-paper-tells-the-full-story/.

[21] Alireza Zandieh, Artificial Intelligence, Economic Inequality, and the Financial Hurdles to Sustainable Peace: Navigating the Interconnected Challenges of the 21st Century, 13 SOC. SCIS. & HUMANS. OPEN 102687 (2026), https://doi.org/10.1016/j.ssaho.2026.102687.

[22] Fatmir Besimi, The Political Economy of AI: Productivity, Power and the New Social Contract, STRATEGERS (May 24, 2026), https://www.strategers.org/the-political-economy-of-ai-productivity-power-and-the-new-social-contract/.

[23] Anton Korinek, Martin Schindler & Joseph E. Stiglitz, Technological Progress and Artificial Intelligence, in How to Achieve Inclusive Growth 163 (Valerie Cerra et al. eds., 2021), https://doi.org/10.1093/oso/9780192846938.003.0005.

[24] Daron Acemoglu & Pascual Restrepo, Automation and New Tasks: How Technology Displaces and Reinstates Labor, 33 J. ECON. PERSPS. 3 (2019), https://doi.org/10.1257/jep.33.2.3.

[25] UNESCO, Can Technological Progress Build Prosperity? (May 4, 2026), https://www.unesco.org/en/articles/can-technological-progress-build-prosperity.