The Innate Economic Preferences of Language Models

2026/07/28 by Joy Buchanan, Joshua Foster
Economics, Econometrics and Finance · #econ.EM

paper · pdf

arxiv created 2026/07/28 · arxiv updated 2026/07/30

Abstract

Language models increasingly settle real resource tradeoffs on behalf of principals yet their economic preferences remain unobserved. We demonstrate their generation rule is isomorphic to the random utility model of discrete choice. This allows internal logit scores to structurally identify preferences. Estimating risk attitudes across twelve models in a portfolio task reveals universal but heterogeneous risk aversion. Although models reject strictly dominated options, their elicited preferences fail invariance tests and violate the independence of irrelevant alternatives across varying experimental prompts. Finally, fine tuning establishes that a principal can explicitly engineer a target risk attitude.

Citations

• Aligning Large Language Model Agents with Rational and Moral Preferences: A Supervised Fine-Tuning Approach
• Evaluating and Aligning Human Economic Risk Preferences in LLMs
• The Order Effect: Investigating Prompt Sensitivity to Input Order in LLMs
• Can an LLM Learn Preferences from Choice Data?
• Large Language Models as Simulated Economic Agents: What Can We Learn from Homo Silicus?
• Training language models to follow instructions with human feedback
• BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
• LoRA: Low-Rank Adaptation of Large Language Models
• Language Models are Few-Shot Learners
• The Curious Case of Neural Text Degeneration
• Algorithm appreciation: People prefer algorithmic to human judgment
• Deep reinforcement learning from human preferences
• Algorithm aversion: People erroneously avoid algorithms after seeing them err.
• Constructing Efficient Stated Choice Experimental Designs
• Estimating Discrete-Choice Models of Product Differentiation
• Maximum score estimation of the stochastic utility model of choice
• The Construction of Utility Functions from Expenditure Data

Related

• Can an LLM Learn Preferences from Choice Data?
• Evaluating and Aligning Human Economic Risk Preferences in LLMs
• Generalized Subjective Lexicographic Expected Utility Representation
• The Emergence of Economic Rationality of GPT
• Can Large Language Models Capture Human Risk Preferences? A Cross-Cultural Study