AI Socialism: Who Will Own the Automated Future?

Summary

Robert Chapman examines Zhao Levi’s evolving argument that opposition to the AI industry should be paired with opposition to capitalist ownership, while stressing that Levi’s proposed DSA AI Program Working Group is an invitation to debate rather than an adopted DSA position. The essay argues that DSA’s existing program addresses common goods, unions, public ownership, layoffs and surveillance but does not yet explain how AI should be developed, governed and shared. It proposes “dual power”: unions and communities should resist harmful uses such as uncompensated training, algorithmic management and surveillance, while cooperatives and public institutions build AI services people can choose. Examples include worker-governed clinical tools, public AI tutors, shared energy software and publicly financed research labs, though the author presents these as pilots rather than existing institutions. The proposed system would need procurement, cooperative ownership, independent evaluation, controlled access to dangerous capabilities and transparent reporting of costs, performance and failures. The essay is wary of a centralized intelligence infrastructure, arguing that public ownership alone does not guarantee dissent, privacy, accountability or the ability to leave. Drawing on council communism, syndicalism and Chile’s Project Cybersyn, it advocates workplace councils, recallable delegates, civic juries, sectoral bargaining and public-interest oversight, with affected users represented alongside workers. AI could help coordinate production and energy demand, but people would set rights and goals and retain the power to contest decisions. The author also proposes transitional energy vouchers while acknowledging that energy cannot measure all forms of value and that essential services should remain guaranteed. The article cites an International Energy Agency projection that data-center electricity use could exceed 945 terawatt-hours globally by 2030, with AI the largest driver, and calls for disclosure, community consent, labor standards and payment for grid upgrades. Its near-term proposal is a universal AI dividend, or “Social Security for All,” funded through public equity, taxes, royalties or returns from public infrastructure without replacing existing benefits and services. The concluding agenda combines worker power, a public and cooperative AI commons, a dividend, accountable clean-energy expansion, civic oversight and a transition beyond wage dependence, while leaving military use, surveillance, privacy, discrimination, creative consent and dangerous capabilities as issues for further debate.