Much of today’s policy debate over artificial intelligence (AI) revolves around instruments designed to change the costs of access to AI; this report describes a formal model of AI adoption and proliferation that focuses on the demand for AI capability.

Buying Time Against AI Proliferation

The Economics of Governed Access

Tobias Sytsma

ResearchPublished Sep 3, 2026

Much of today’s policy debate over artificial intelligence (AI) focuses on efforts to change the costs of access to AI—to shift relative costs such that AI adoption is steered toward channels that society can monitor and away from channels it cannot.

Policy analysis has concentrated on the supply side of this problem, with such mechanisms as export restrictions, licensing requirements, and controls on computing hardware. But where AI capabilities end up depends on the choices of organizations and individuals about what to adopt and through which channels, and it is on this demand side that cost-shifting instruments operate. This demand side of the problem has received far less systematic analysis.

This report addresses this gap with a framework for judging when cost-shifting steers adoption as intended, when it accomplishes little, and when it backfires and makes proliferation worse. The author details a formal model in which AI adoption is driven by relative costs across access channels, and governance instruments operate by manipulating those relative costs. The central question is: When policy changes the relative cost of AI access, how do users reallocate across governed and ungoverned channels, and under what conditions does that reallocation reduce, fail to affect, or accelerate uncontrolled proliferation of AI?

Key Takeaways

Compliance burdens on governed AI access channels can make uncontrolled AI proliferation worse

  • Raising the cost of controlled frontier access without corresponding pressure on ungoverned alternatives can redirect user adoption toward ungoverned access channels. In simulations conducted by the author, this occurred in roughly one-quarter of relevant scenarios.

Whether AI access governance can sufficiently delay or suppress ungoverned access growth depends on three demand-side conditions

  • Three indicators predict whether policies can suppress ungoverned AI access: (1) how responsive adopters are to cost differences across channels, (2) how effectively policy can raise the cost of ungoverned AI access, and (3) how costly it is for users to switch between access channels.

AI access governance policies are no longer able to shift users toward governed access channels once AI capabilities are sufficiently strong

  • Once AI capabilities become sufficiently valuable to users, policies that attempt to shift the relative costs of governed vs. ungoverned AI access lose their traction. Once AI capabilities advance far enough, no policy evaluated in the modeling framework can suppress ungoverned access.

The distinction between AI capability dynamics and access-cost dynamics matters for governance of the technology

  • A framing centered on capability thresholds implies that governance should track and respond to specific technical milestones; a framing centered on price dynamics implies that governance should track and respond to the cost of access. The latter may be more observable, more amenable to policy intervention, and more predictive of who adopts what and through which channels.

Topics

Document Details

Citation

Chicago Manual of Style

Sytsma, Tobias, Buying Time Against AI Proliferation: The Economics of Governed Access. Santa Monica, CA: RAND Corporation, 2026. https://www.rand.org/pubs/research_reports/RRA4996-1.html.
BibTeX RIS

Research conducted by

This research was independently initiated and conducted by the Center for the Geopolitics of Artificial General Intelligence within RAND Global and Emerging Risks using income from operations and gifts from RAND supporters.

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.