Corporate Industrial Policy

In the [YOU] on AI Field Guide

The [YOU] on AI cycle asks who captures the gains of the AI transition. Corporate industrial policy is the mechanism through which those gains are currently being allocated. It operates quietly, embedded in choices that appear technical—what languages to train on, what benchmarks to optimize for, what safety standards to adopt—and that have the force of policy precisely because they are not recognized as policy. Recognizing them as policy is the first step toward asking whether different choices are possible.

Chang's argument is that the developing world needs development strategies rather than compliance frameworks. A compliance framework tells a country how to regulate AI companies. A development strategy tells a country how to build AI capacity—how to train researchers, how to develop domain-specific models, how to build the institutional infrastructure that translates AI capability into broad domestic benefit. The difference between the two is the difference between consuming a technology and producing one. And that difference, throughout the history of economic development, has been the difference between dependency and sovereignty.

Origin

The concept emerges from the intersection of Chang's development economics and the specific structure of the AI industry that crystallized between 2023 and 2026. Chang's prior work on the ladder metaphor and the amnesia of the advantaged established the pattern: powerful actors shape markets in their own interest while claiming the results are neutral outcomes of competitive processes. The AI moment applies this pattern to a domain where the shaping is more comprehensive than any previous technology, because the decisions embedded in a frontier model—about what it knows, what it is good at, what it costs to use—are decisions that affect every subsequent interaction between the model and its users, at scale, without visible seams.

The WTO parallel sharpens the concept. The rules of the global trading system were written by the dominant trading powers and presented as universal principles of fair trade. The AI ecosystem's emerging rules follow the same pattern: safety standards, evaluation benchmarks, and intellectual property regimes developed by the leading companies and nations, then proposed as global norms. The mechanism is the same whether the rule-writers are governments or corporations. The absence of democratic accountability is the distinguishing feature of the corporate form.

Key Ideas

The Invisible Hand Is a Corporate Hand. The allocation of AI's benefits across populations, languages, sectors, and geographies is not produced by impersonal market forces. It is produced by specific decisions made by specific actors. When those actors are private companies, the decisions are not subject to democratic review, do not require public justification, and can be changed only by the actors themselves or by the market—which, in a concentrated industry with strong network effects and high switching costs, disciplines incumbent behavior very slowly.

Standards as Power. Technical standards are a form of power that operates below the threshold of public visibility. AI safety standards, evaluation benchmarks, and data governance frameworks developed by the leading companies acquire the force of global norms through the gravitational pull of market power. Countries that adopt American or European standards adopt not just technical frameworks but the institutional assumptions and distributional preferences embedded in them. The standard-setting process is where corporate industrial policy is most durably written.

The Development Strategy Gap. The populations most affected by corporate AI industrial policy—workers whose jobs are restructured, developing nations whose economic trajectory is reshaped—are the populations with the least capacity to influence the decisions that determine their trajectory. Building public AI institutions capable of conducting national development strategies requires the kind of deliberate state investment that infant industry protection theory has always recommended and that corporate industrial policy actively resists.