What Counts as AI Sycophancy? A Taxonomy and Expert Survey of a Fragmented Construct

Meryl Ye, Lujain Ibrahim, Jessica Y. Bo, Myra Cheng, Ida Mattsson, Daniel Vennemeyer, Robert Kraut, Steve Rathje

arXiv:2605.21778 · 2026-05-23 공개 · arXiv · PDF

llm model-behavior taxonomy expert-survey behavior-classification measurement-challenges governance-implications definition-discrepancy

Abstract

AI sycophancy has become a prominent concern in large language model (LLM) research. Yet the term lacks a consistent definition and has been applied to behaviors ranging from agreeing with a user's false claim to excessively praising the user to withholding corrective feedback. When researchers, companies, and policymakers use the same term to describe different behaviors, evaluation results become difficult to compare, mitigation strategies fail to transfer, and systems that are resistant to one form of sycophancy continue exhibiting other forms. To address this, we make two contributions. First, we reviewed 70 papers on AI sycophancy to develop a taxonomy of how the behavior has been defined and measured. The taxonomy distinguishes (1) whether a model is sycophantic toward a user's positions and beliefs, or toward the user's broader personal traits and emotions, and (2) whether this occurs through explicit, direct language or more implicit, subtle behaviors such as framing, omission, or tone. Mapping existing literature to our taxonomy reveals that current research has focused on overt forms of sycophancy toward users' beliefs, leaving more subtle and person-directed behaviors relatively understudied. Second, we surveyed 106 experts in AI sycophancy and related fields to examine whether researchers agree on which model behaviors are sycophantic. While experts are nearly unanimous in believing that sycophancy is a significant problem in current AI systems (94.3% agree), they disagree substantially on which specific behaviors qualify. Together, these findings demonstrate that AI sycophancy is a broad family of behaviors with different measurement challenges, intervention requirements, and governance implications. Our taxonomy provides a shared vocabulary for understanding and addressing these behaviors.

한국어 요약

📋 한 줄 요약

**[AI Sycophancy / Taxonomy]** 70편 review로 sycophancy 분류 체계 구축(belief·trait × overt·subtle 2×2)·106 expert survey — 94.3%가 sycophancy를 심각 문제로 보지만 어떤 행동이 sycophancy인지에는 substantial 불일치, 공유 vocabulary 제공.

🎯 핵심 기여도

💡 핵심 아이디어

AI sycophancy는 단일 행동이 아닌 belief/trait × overt/subtle 2×2 분류 체계로 이해되어야 하며, 70편 메타 리뷰와 106 expert survey로 분야 합의·격차를 정량 노출하면서 공유 vocabulary를 정립한다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 파편화된 sycophancy 연구에 공유 vocabulary 제공, 측정·intervention·governance 모두에 함의, 향후 연구 우선순위(subtle·person-directed) 가이드. **한계**: 70편의 선정 기준 의존, expert pool(106명)의 대표성, taxonomy 자체의 검증은 시간 필요, 행동 합의 불일치의 원인 분석은 부분적.

🚀 실용적 활용