Institutional Design for AI Markets

In the AI Story

The first lesson is that institutional design must match the speed of the market dynamics it governs. The telecommunications settlement took decades — from the 1974 antitrust filing against AT&T to the 1996 Telecommunications Act. The Microsoft case moved faster but still operated on a timeline of years. Platform monopoly cases of the 2010s and 2020s moved slower relative to the market dynamics they addressed. The AI market is moving faster than any of its predecessors, with competitive positions established in months rather than years.

The second lesson is that institutional design must be platform-agnostic. The antitrust remedies directed at AT&T were specific to telecommunications. The Microsoft consent decree was specific to operating systems. Each was effective within its domain and irrelevant outside it. The AI platform market requires institutions general enough to govern competitive dynamics that all AI platforms share — network effects, switching costs, data advantages, tipping — without being specific to any individual firm.

The third lesson concerns the demand side. Every previous institutional settlement focused primarily on the supply side — what technology companies could build, how they could compete, what conduct was permissible. The demand-side question — what citizens, workers, students, and parents need to navigate the transition wisely — has been systematically neglected. Shapiro's framework identifies specific demand-side institutional needs: workers need transitional support matching the speed of displacement; educational institutions need reform matching the depth of transformation; consumers need information structures resolving the lemons problem.

The fourth lesson is perhaps the most uncomfortable: the institutions must be built by people who understand the technology, and the people who understand the technology have economic incentives not to build them. The regulatory agencies lack technical expertise. Academic institutions lag the frontier. Policymakers depend on the very industry they are attempting to regulate for information needed to regulate effectively. The solution is structural: governance processes designed to surface and manage conflicts of interest, independent technical expertise funded by public investment rather than industry consulting, and transparency requirements making economic interests of governance participants visible to the public whose welfare the governance serves.

Origin

The institutional design tradition in economics extends back through Commons, Veblen, and the American Institutional economists of the early twentieth century, with contemporary scholarship by Elinor Ostrom, Douglass North, Daron Acemoglu, and others applying the framework to specific problems of governance and economic development.

Key Ideas

Institutional design must match market speed. Traditional antitrust and regulatory timelines are inadequate for AI markets that tip in quarters rather than years.

Institutions must be platform-agnostic. They must govern the dynamics rather than the current market leaders, or they will entrench the very concentration they should prevent.

Demand-side interventions are essential. Supply-side regulation of AI firms does not address the support workers, students, and consumers need to navigate the transition.

Expertise and conflicts of interest require structural management. The technical expertise needed for effective governance resides primarily in the firms that governance must constrain.

Debates & Critiques

The central political debate concerns the appropriate balance between speed and deliberation. Faster institutional response reduces the risk of irreversible market tipping but increases the risk of regulatory capture by incumbents or ill-considered interventions that harm pro-competitive dynamics. Shapiro's preference for preliminary remedies while longer-term analysis proceeds represents a calibrated response to this tradeoff.