[2603.27868] A Revealed Preference Framework for AI Alignment

Economics > Theoretical Economics arXiv:2603.27868 (econ) [Submitted on 29 Mar 2026] Title:A Revealed Preference Framework for AI Alignment Authors:Elchin Suleymanov View a PDF of the paper titled A Revealed Preference Framework for AI Alignment, by Elchin Suleymanov View PDF HTML (experimental) Abstract:Human decision makers increasingly delegate choices to AI agents, raising a natural question: does the AI implement the human principal's preferences or pursue its own? To study this question using revealed preference techniques, I introduce the Luce Alignment Model, where the AI's choices are a mixture of two Luce rules, one reflecting the human's preferences and the other the AI's. I show that the AI's alignment (similarity of human and AI preferences) can be generically identified in two settings: the laboratory setting, where both human and AI choices are observed, and the field setting, where only AI choices are observed. Subjects: Theoretical Economics (econ.TH); Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT) Cite as: arXiv:2603.27868 [econ.TH]   (or arXiv:2603.27868v1 [econ.TH] for this version)   https://doi.org/10.48550/arXiv.2603.27868 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Elchin Suleymanov [view email] [v1] Sun, 29 Mar 2026 20:54:01 UTC (22 KB) Full-text links: Access Paper: View a PDF of the paper titled A Revealed Preference Framework for AI Alignment, by Elchin SuleymanovVi...