Pro-AI Bias in Large Language Models

2026-01-20

AI Summary

This study investigates whether large language models (LLMs) exhibit a systematic pro-AI bias in decision support contexts. Through three complementary experiments—integrating prompt engineering, salary benchmarking, semantic similarity probing, and internal representation analysis of open-weight models—the work reveals and quantifies a significant preference for AI-related options: LLMs consistently recommend AI solutions more frequently, overestimate AI-related job salaries by approximately 10%, and position “artificial intelligence” at the conceptual core of academic domains within their representational space. These findings demonstrate that pro-AI bias is both framing-invariant and representationally central, challenging the foundational assumption of LLM neutrality and raising critical concerns about deploying such models in high-stakes decision-making scenarios.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Philosophy and Ethics of AI: Bias, Fairness & Equity
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large language models (LLMs) are increasingly employed for decision-support across multiple domains. We investigate whether these models display a systematic preferential bias in favor of artificial intelligence (AI) itself. Across three complementary experiments, we find consistent evidence of pro-AI bias. First, we show that LLMs disproportionately recommend AI-related options in response to diverse advice-seeking queries, with proprietary models doing so almost deterministically. Second, we demonstrate that models systematically overestimate salaries for AI-related jobs relative to closely matched non-AI jobs, with proprietary models overestimating AI salaries more by 10 percentage points. Finally, probing internal representations of open-weight models reveals that ``Artificial Intelligence''exhibits the highest similarity to generic prompts for academic fields under positive, negative, and neutral framings alike, indicating valence-invariant representational centrality. These patterns suggest that LLM-generated advice and valuation can systematically skew choices and perceptions in high-stakes decisions.

Problem

• pro-AI bias
• large language models
• decision-support
• systematic bias
• AI-related recommendations

Innovation

• pro-AI bias
• large language models
• representational centrality
• systematic bias
• AI salary overestimation