Demand for a Public AI Option

Abstract

We study how entry by a public AI option would reshape demand for incumbent private services. Given that large-scale consumer-facing public AI provision does not yet exist, we leverage a discrete choice experiment with 4,215 market-eligible U.S. adults to provide an ex-ante evaluation of public option entry. We find that public option availability sharply compresses incumbent residual brand valuations and increases private price sensitivity. Within the public-option regime, a sponsored/co-branded implementation attracts higher demand than direct government provision, holding displayed attributes fixed. Counterfactual simulations illustrate partial-equilibrium trade-offs between competitive pressure and take-up under varying user fees.

Keywords: Artificial Intelligence, Public Provision, Discrete Choice Experiment

JEL Classification: D12, D91, H42, O33