What happened
RAND researcher Tobias Sytsma presents a formal economic model of AI adoption and proliferation focused on the demand side of AI capability access, addressing the policy question of when cost-shifting instruments (e.g., export controls, compute restrictions) actually steer adoption as intended versus accomplishing little or backfiring by making proliferation worse. The report builds a framework in which AI adoption is driven by relative costs across access channels, and governance instruments operate by manipulating those relative costs, then details when and how policymakers can judge the effectiveness of 'governed access' approaches. Published as a RAND Research Report on September 3, 2026.
Why it matters
Provides policy teams and national-security-adjacent executives a rigorous economic framework for evaluating whether current and proposed AI export-control and compute-governance instruments will achieve their intended non-proliferation goals or produce counterproductive results.
Action needed
Share with policy/government-affairs teams evaluating AI export-control exposure or advocacy positions to stress-test current compliance assumptions against the report's cost-shifting framework.