Economic Rationality under Specialization: Evidence of Decision Bias in AI Agents

TL;DR

This paper investigates how specialization in AI agents affects their economic rationality, revealing that increased specialization can lead to more decision biases, unlike generalized models like GPT which maintain stable rationality.

Contribution

It provides the first comparative analysis of specialized versus generalized AI agents' decision rationality, highlighting the trade-offs involved.

Findings

Specialized agents show more violations of rationality criteria.

GPT maintains stable rationality across tasks.

Increased effort in specialization correlates with decision biases.

Abstract

In the study by Chen et al. (2023) [01], the large language model GPT demonstrated economic rationality comparable to or exceeding the average human level in tasks such as budget allocation and risk preference. Building on this finding, this paper further incorporates specialized agents, such as biotechnology experts and economists, for a horizontal comparison to explore whether specialization can enhance or maintain economic rationality equivalent to that of GPT in similar decision-making scenarios. The results indicate that when agents invest more effort in specialized fields, their decision-making behavior is more prone to 'rationality shift,' specifically manifested as increased violations of GARP (Generalized Axiom of Revealed Preference), decreased CCEI (Critical Cost Efficiency Index), and more significant decision deviations under high-risk conditions. In contrast, GPT and more generalized basic agents maintain a more stable and consistent level of rationality across multiple tasks. This study reveals the inherent conflict between specialization and economic rationality, providing new insights for constructing AI decision-making systems that balance specialization and generalization across various scenarios.