Abstract
This paper investigates how generative AI affects economic valuation and market efficiency through a controlled laboratory experiment (N = 264) featuring a real-effort Principal-Agent task. Using an incentive-compatible Becker-DeGroot-Marschak mechanism, we find that AI availability acts as a powerful market lubricant: Principals significantly increase their willingness to pay (WTP) driven by rational expectations of higher output quality, while Agents lower their willingness to accept (WTA) due to reduced effort costs. These dual shifts expand the bargaining zone, causing contract matching rates to surge from 65.1% to 96.9% and nearly doubling the total economic surplus. However, we also document a robust behavioral friction: Compared to the price of acquiring AI access, agents endowing with AI access demand a 26% premium to relinquish it. These findings demonstrate that while generative AI enhances market efficiency through productivity gains, it is rapidly internalized as a psychological entitlement, creating significant endowment effects that may complicate technology adoption.