Whose AI? Whose Abundance?
A Workshop Reader II
Prepared for the UFT-CWE Chapter
Comments and questions welcome — please write to xamyxamu@hotmail.com
A note before you begin. This is not an inventory of every name in the AI discourse. It is an argument made through the figures and movements the argument requires. Read it as such. The four parts move from the debate as it is publicly conducted, to the question that debate largely declines to ask, to the alternative already operating at scale, to the political consolidation now actively foreclosing that alternative. Each part names figures most readers will recognize and figures most will not. The unfamiliar ones are usually load-bearing.
The document is a working text, not a reference. Add to it. Argue with it. Bring it into the rooms where you teach, organize, study, or work. The People's Share proceeds by census of particular people and particular things rather than by ideology, and this reader operates on the same principle: these specific figures, doing this specific work, in this specific moment.
I.The AI Debate as It Is Currently Conducted
The figures in Part I disagree fiercely with one another, but most of them share a frame. The frame is the technology itself — what it can do, what it cannot do, when it will do more, what dangers attend its capabilities, what safeguards might constrain them. The debate is conducted by researchers, founders, ethicists, regulators, and journalists, and it occupies nearly all of the oxygen in the public conversation about AI.
Part I presents this debate on its own terms, organized into seven movements: existential warning; the figures who argue the paradigm itself is conceptually incomplete; the critics of present and documented harm; the leaders building and accelerating the technology; the maximalists who hold the transformation is near; the alignment researchers working the engineering problem; and the accountability reporters who form the hinge into Part II. That shared frame is broken from inside along two independent seams — one epistemological, where the paradigm critics argue the method does less than its builders claim, and one political, where the accountability reporters make the ownership question unavoidable. The document follows the political seam into Part II; the epistemological one stands on its own.
Read Part III.Who Owns It?
Part I asks whether AI will be safe, fair, aligned, or beneficial. Part II asks who owns the machines, who owns the data, and who decides — and treats those questions as logically prior to the questions Part I conducts. The Autotroph argument that organizes The People's Share applies here directly: the distributional question is logically prior to the productivity question, because the abundance a technology generates flows to whoever owns the productive capacity, regardless of how abundant the technology becomes. A technology can be perfectly safe and entirely fair in its outputs and still concentrate wealth and power radically, if its ownership is concentrated.
Part II is organized in three movements: the economists and political theorists naming the problem at the level of political economy; the figures developing alternative governance models for data, compute, and platform infrastructure; and the post-capitalist theorists designing what democratic ownership of AI would actually look like in operation. The section moves from diagnosis to design, and ends — at the threshold of Part III — with the figures who insist the question is not theoretical because the alternative is already operating.
Read Part IIIII.Proof, Not Aspiration
The cooperatives in Part III are not thought experiments. They employ hundreds of thousands of workers, generate billions in annual revenue, and have operated democratically — under one-worker-one-vote governance, with wage ratios that would be illegal in most US corporations only because they are too equal — for decades. They are present in every sector of the economy and in dozens of countries. The global cooperative sector encompasses roughly three million cooperative enterprises with more than one billion individual members; the largest three hundred cooperatives generate combined annual turnover above two trillion dollars. The figures come from the International Cooperative Alliance, and they almost never appear in mainstream economic discourse.
The question Part III answers is not whether democratic ownership of productive infrastructure can work. It is why, given that it manifestly does, it remains marginal — and what would be required to extend it to the next general-purpose technology before enclosure completes.
Read Part IIIIV.Consolidation
The technology described in Part I is not developing in a political vacuum, and the alternative described in Parts II and III is not foreclosed by technical necessity but by political project. Part IV names the project. The figures in this section are not interchangeable. The ideologues supply doctrine. The operators build infrastructure and staff policy. The state actors deploy. Their political alignments differ; some are open rivals to one another. What they share is a structural relationship to AI as the principal technology of the coming decade and a willingness — in some cases an explicit program — to organize political power around its concentration rather than its distribution.
The receipts in Part IV are direct quotations and signed manifestos. The figures convict themselves.
Read Part IVThe framing the consolidation has worked hardest to establish — that the choice is between technological progress and democratic governance, that to slow AI is to fall behind China, that the firms building the technology are the only ones competent to govern it — is the consolidation's principal rhetorical achievement, and it is not true. The choice is not between progress and democracy. The choice is between two different political projects organized around the same productive infrastructure. One project — democratic ownership, cooperative governance, the commons — is operating, at scale, in dozens of countries, for decades, generating billions in revenue, employing hundreds of millions of people, distributing the gains of productivity to those who produce them. The other project is being assembled, in real time, in front of us, by a small number of named figures with stated political programs, funded by a smaller number of sovereign and corporate sources, implemented through a tightening alignment between private digital infrastructure and state power.
The Autotroph argument is that the distributional question is logically prior. The People's Share is the proposition that the public — the actual public, not its representatives or its avatars — owns a share of the productive capacity, including the productive capacity called artificial intelligence, by right and not by gift. The figures in this reader are some of the people whose work makes that argument legible. The argument itself belongs to the room.
Read this document. Argue with it. Use it.