Why the people building the AI era should own a share of it
People should own a stake in the systems they help make valuable.

A newspaper chain in Palo Alto

In 1956, employees of Peninsula Newspapers in California acquired the company through an arrangement developed with Louis Kelso. It became a foundational example in the history of the Employee Stock Ownership Plan. The practical question was how workers could acquire an interest in a productive business without already having the wealth to buy it outright.

Kelso’s argument was that the concentration of productive capital limited who could benefit from economic growth. Wages compensated labor; ownership supplied a different claim on the enterprise. He sought mechanisms through which workers could acquire that claim without first becoming wealthy.

The Peninsula Newspapers situation gave him a place to test the theory. The employees did not have the money to buy the company. The company itself, however, had future earnings, and those earnings would belong to whoever owned the company. If the employees could borrow against the future earnings — and then use the future earnings to repay the loan, gradually, over years — they could become the owners of a company they had not been able to afford on the day of the purchase. So Kelso designed the structure to do it: an Employee Stock Ownership Trust, organized to hold company stock on behalf of the employees, financed through a leveraged loan and repaid out of company earnings. The exemption it needed was new; Kelso secured it; the trust was established; and in 1956 the employees of Peninsula Newspapers, Inc., became, over time, the owners of the company they had built. It operated as employee-owned for the next quarter century, paying out millions in benefits to participants, before eventually being acquired. By then the form Kelso had designed had begun to spread.

Kelso developed the ownership argument with Mortimer Adler in The Capitalist Manifesto, published in 1958. His later work with Senator Russell Long helped advance federal recognition of ESOPs. The history matters here because a proposal required both a financing mechanism and sustained institutional work.

ESOP provisions entered the federal retirement-law framework through the Employee Retirement Income Security Act of 1974 and subsequent legislation. The resulting institution is a regulated retirement plan, with requirements that do not automatically extend to a new system of AI contribution or ownership.

What Kelso and Long built between them, over roughly two decades of legal and political work, was a form. The form is still here. The argument of this chapter is that it is the third of the institutional ancestors of Cyberwealths, and that the work the AI era now requires — building broad-based ownership for the productive capacity that has been quietly accumulating for the past three decades — is the work the form has spent a lifetime preparing to do.

What an employee-owned enterprise is, and what its discipline costs

Employee ownership includes several different arrangements. In an ESOP, a retirement-plan trust holds company shares for participants, with accounts and distributions governed by plan terms and law. Benefits may be distributed in cash rather than continuing company shares. A worker cooperative uses a different ownership and governance structure, commonly based on democratic member control. Employee ownership does not, by itself, mean employees manage the business or vote on every decision.

These arrangements can broaden the claim workers have on the value they help create. Some companies are entirely employee-owned; others combine employee and outside ownership. The important questions are how much is owned, what rights accompany it, how benefits are allocated, and who exercises control.

Broad ownership brings responsibilities as well as benefits: financing, valuation, succession, liquidity, and obligations when participants leave or retire. Democratic participation requires time and competence where it is part of the model. None of these arrangements guarantees slower growth, consensus, or better management.

Research summarized by the National Center for Employee Ownership reports favorable associations between employee ownership and several worker and business outcomes. Results depend on the study, comparison group, and form of participation; ownership should not be treated as a single proven cause of every advantage. The practical case combines a share in value with working conditions and governance that make the share meaningful.

The form has many faces

Some sense of the form’s range is worth carrying into the rest of the chapter, because the AI-era institutions the trilogy argues for need to draw on as much of the form’s accumulated wisdom as possible.

The ESOP is an established American route to employee ownership across many industries. It can hold a minority interest or all of a company’s stock. Its prevalence should not obscure the distinction between participation in a retirement plan and democratic governance of a workplace.

American worker cooperatives include Equal Exchange and Cooperative Home Care Associates. Their democratic ownership should be distinguished from an ESOP. King Arthur Baking, another familiar employee-owned business, belongs in the ESOP discussion rather than being listed as a worker cooperative.

The Mondragon cooperative movement grew from the first industrial cooperative established in 1956 in Spain’s Basque region, with the influence of José María Arizmendiarrieta and the people he taught and organized. Its later network demonstrates the range cooperative institutions can reach, as well as the complexity of sustaining member governance across different businesses and jurisdictions.

We have been the productive capacity

The familiar claim is that people who help a business create value can also hold an ownership interest in it. AI complicates the boundaries of contribution. Paid training work, licensed material, voluntary participation, and personal records involve different relationships. They should not all be treated as equivalent, and not every digital trace is necessarily used to train a model.

AI training can draw on human-created text, images, code, demonstrations, and evaluations. The terms differ across datasets and services. Some contributions are paid or licensed; others raise contested questions about permission and compensation. A future ownership design needs to make the relationship explicit before participation.

Human contribution is part of AI’s productive capacity. The question is what claims and choices contributors should have in the institutions that develop and use it.

Useful services can create real reciprocal benefits while still leaving questions about secondary uses, ownership, and bargaining power unresolved. Acknowledging the benefit should not prevent examination of the terms.

The ESOP precedent shows one way to connect work with a claim on productive capital. It does not establish how every contribution to an AI system should be valued. That requires a new account of eligible contributions, rights, and benefits, with participation by the people affected.

This chapter is not arguing that platforms owe their users back pay for thirty years of unrecognized contribution. That argument would be both impractical and conceptually wrong — back pay is a wage claim, and what was produced was something other than wages. What the chapter argues is that the AI era now being built requires, by the same institutional logic that produced the ESOP for Peninsula Newspapers, ownership architecture that includes the people whose continuing contribution makes the era valuable. The architecture of the past thirty years cannot be unbuilt. The architecture of the next thirty years can be built differently.

From payment to a continuing stake

Public benefits and ownership claims are different arrangements. A public benefit may be owed under law and financed through taxes, dedicated contributions, or other public resources. Its durability depends on institutions and political decisions. That does not make it a discretionary favor or imply that it lacks value.

Ownership can provide a claim on value and, depending on the instrument, governance rights. It can also expose the holder to losses, dilution, illiquidity, or an enterprise’s failure. Property and corporate rights depend on law too. A stake is not automatically perpetual, inheritable, liquid, or independent of politics.

Universal basic income proposals seek broad, regular payments under public arrangements; their eligibility and financing vary. They address a different question from direct ownership of productive assets. A society could pursue income security and broader ownership together. My emphasis here is what people can hold and govern, not a claim that income support must be abandoned.

The employee-ownership tradition adds a question to debates about income: can people hold a share of productive systems and help govern them? The answer need not displace public benefits. It can broaden the institutions through which people gain security and a voice.

From employment to contribution

An ESOP operates within an employment and retirement-plan framework. AI contribution can involve other relationships: a learner’s authorized record, a teacher’s licensed materials, a resident’s civic knowledge, or a worker’s paid evaluation. Their terms differ. Extending ownership beyond employees requires an explicit new arrangement, not an assumption that all participation already produces a claim on model equity.

The arrangement needs an agreed way to recognize contributions, allocate stakes, and resolve disputes. Its terms should define the contributors, assets, governance, and benefits before anyone is asked to exchange work or knowledge for ownership.

This is what it means to extend the form from employment to contribution. The Employee-Owned Enterprise is the ancestor; the Cyberwealths Enterprise — the AI-era equivalent — is what the form becomes when contribution is recognized as the entry point into ownership rather than only formal employment. The same institutional grammar applies. There is still a productive system; there are still contributors whose work makes it valuable; there should still be a mechanism by which the contribution becomes an ownership stake. The grammar simply widens to admit contribution that does not arrive through a W-2.

This widening is not a license for vagueness. The forms of contribution that count for ownership have to be specified, recorded, governed, and audited — exactly as the ESOP specifies, records, governs, and audits formal employment. Authored experience contributed under consent to a Mutual AI Credit Union, model-feedback provided in the course of using AI tools, learning records built into community education infrastructure, civic data provided to local knowledge graphs, AI representative interactions collected to improve shared models — each of these can be specified with the same legal-and-administrative precision the ESOP has used for formal employment. The institutional work is significant but not exotic; it is the kind of work the cooperative tradition has been doing, in different applications, for more than a century. The phrase the trilogy uses for this widening is society as the enterprise. It does not mean society becomes one large company, or that every person becomes an employee of a collective corporation. It means the productive base of the AI era is social — generated by the contributions of millions of people, often through ordinary use of ordinary services, often without anyone realizing the contribution was being made — and that the ownership architecture has to reflect the structure of its productive base. The base is social. The ownership has to be social too.

The data backpack as ownership account

The previous chapter introduced the data backpack as the member account of a Mutual AI Credit Union, the AI-era equivalent of a credit union share account. The Employee-Owned Enterprise adds a second function to the same instrument. The data backpack is also the ownership account of a Cyberwealths Enterprise. This is not a different account; it is the same account doing two functions, in the way that an ESOP plan account is, for an employee, both a record of accrued ownership in the company and a vehicle for retirement savings. The backpack records the member’s contributions — to shared models, to community knowledge, to civic data, to authored experience available under consent for AI training — and it records the member’s ownership stake in the Cyberwealths Enterprise those contributions help to constitute. The two readings are connected by an explicit logic: contribution is what builds the stake; the stake is what the contribution is exchanged for.

A member could accumulate an agreed stake rather than receive only a one-time fee. The arrangement would need to specify voting rights, distributions, valuation, transfer, inheritance, and exit. Those terms should be understandable before contribution occurs. Equal membership voice need not depend on the size of a financial account.

The Cyberwealths Enterprise would operate shared AI services with an ownership arrangement tied to defined contributions and membership. It would need separate rules for the economic stake and for personal-data permissions. Leaving the enterprise, redeeming a stake, exporting records, and withdrawing future authorization are different actions, and some past uses may not be reversible.

These enterprises can develop alongside existing platforms, providing another way to organize AI services and share their value. Their reach will depend on adoption, competitive services, workable agreements, and the public institutions through which new ownership arrangements gain recognition.

What the Cyberwealths Enterprise inherits from Peninsula Newspapers

The employee-ownership tradition provides useful mechanisms to study: trusts, individual accounts, allocation rules, valuation, and benefit distribution. Their adaptation to AI participation is not automatic. A proposal must specify which mechanism it uses, whether contributors are employees or another kind of member, and what legal obligations follow.

Privately held, publicly traded, ESOP-owned, and cooperative enterprises can each contain varied ownership arrangements. The comparison should be made through actual rights: who holds equity, who votes, who receives distributions, who bears risk, and what happens on exit. The proposed Cyberwealths Enterprise must answer those questions concretely.

The Cyberwealths Enterprise remains a proposal. Its purpose would be to connect qualifying contributions with meaningful rights and benefits under member governance. Those rights must be specified rather than assumed from the name.

The architecture is not unusual; American employee-owned enterprises do versions of all of it every day. The Cyberwealths Enterprise applies it to a new substrate — harder in some ways, because the legal vocabulary for contribution-based ownership is less developed than for employment-based ownership and the auditing of social contribution is more complex; easier in others, because the cooperative civic culture still understands the form, the financial and legal infrastructure for trust-based ownership is well-developed, and the political case can be made convincingly across the entire ideological spectrum. The Cyberwealths Enterprise is the third pillar of the architecture the trilogy has built. The Land Trust holds the foundational assets — the substrate of community life on which the AI era depends. The Mutual AI Credit Union pools the member accounts — the means by which individual participation becomes collective capacity. The Cyberwealths Enterprise organizes the productive contribution — the work, broadly understood, by which the era’s value is generated and from which the members’ ownership stakes accumulate. Each of the three refuses a specific trade ordinary commerce would otherwise license. Together, they put the people whose lives the era is built around on the ownership side of it.

From check to stake

Kelso and Long did not live to see the AI era. Their work nevertheless leaves a question for it: how can people acquire a meaningful share in productive systems without first possessing the wealth those systems generate? An inherited mechanism is a starting point for that work, not its finished answer.

The population of possible contributors extends beyond a conventional workplace. That makes the adaptation difficult: an ownership rule must identify eligible contributions, avoid coercing people to disclose private experience, and remain governable at the scale of a real community.

The question facing the AI era is not whether productive value will be created. The value is being created now, every day, at scale, by people who do not yet know they are creating it and by people who do. The question is whether ownership of the productive systems that capture that value will accumulate to a small number of platform shareholders, the way the productive systems of the late industrial era accumulated to a small number of corporate shareholders, or whether the ownership architecture will be built — deliberately, before the wealth has fully consolidated — to include the people whose contribution makes the systems valuable. The two paths produce different futures. The first produces a society that has to argue, every electoral cycle, over what fraction of the captured wealth should be redistributed to the people whose lives generated it: the argument is exhausting, the redistribution is partial, the recipients remain recipients, and the productive systems remain owned by someone else. The second produces a society in which the people whose contribution generates the wealth are also owners of some share of the wealth-generating systems, with governance rights, dividends, and a continuing claim on the systems’ success under defined terms. The first is the entitlement path; it is not a bad path, and it may well be part of any actual settlement. The second is the Cyberwealths path. It is the path this trilogy argues should be built.

From check to stake is the phrase I use for a continuing claim on an institution’s value. A payment already received is real value; a stake carries both possibility and risk. The aim is to give contributors an informed choice and, where appropriate, a lasting role in governing what they help build.

The institutional precedent is available. The next task is a concrete arrangement people can examine before joining: what they contribute, what they receive, what they control, and what happens if the enterprise fails.

Sources

The Kelso Institute (kelsoinstitute.org). Primary source on Louis Kelso’s life and work, the 1956 founding of the first ESOP at Peninsula Newspapers, Inc., the 1973 meeting between Kelso and Senator Russell Long, and the historical chronology of the ESOP movement.

Louis O. Kelso and Mortimer J. Adler. The Capitalist Manifesto. Random House, 1958. Primary intellectual source for Kelso’s theory that the long-run problem of capitalism is not private property but its concentrated ownership, and that broad-based capital ownership is the corrective.

National Center for Employee Ownership (nceo.org). The standing reference for current statistics on American employee-owned enterprises, including ESOP counts, participant numbers, plan assets, and research on the performance of employee-owned companies.

The ESOP Association (esopassociation.org). Trade association reference on the contemporary scale and operation of American ESOP-form companies, including governance practice and member services.

Corey Rosen, John Case, and Martin Staubus, Equity: Why Employee Ownership Is Good for Business (Harvard Business School Press, 2005).

Mondragon Corporation (mondragon-corporation.com). Primary source for the international family of worker-cooperative forms, including Mondragon’s 1956 founding by Father José María Arizmendiarrieta in the Basque region of Spain, parallel in date to the founding of the first ESOP.

U.S. Department of Labor, Employee Benefits Security Administration. Regulatory reference on the federal framework governing ESOPs and related employee-ownership trust structures under the Employee Retirement Income Security Act of 1974.

National Center for Employee Ownership, “What voting rights do ESOP participants have?” https://www.nceo.org/employee-ownership-faq/what-voting-rights-do-esop-participants-have

King Arthur Baking, “Why employee ownership matters” (2018): https://www.kingarthurbaking.com/blog/2018/06/05/why-employee-ownership-matters

U.S. Federation of Worker Cooperatives (usworker.coop). Reference on the contemporary American worker-cooperative sector, including member organizations such as Equal Exchange, Cooperative Home Care Associates, and the Cleveland Evergreen Cooperatives.