mlsanityData to 14 Oct 2026

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

arXiv 2605.21778 ↗First posted 20 May 2026cs.AIcited by 16

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 31 of its 77 references that are in the map. First posted on arXiv 20 May 2026.

Builds on

31 of 77 references.

References · 31

  1. 2023Towards Understanding Sycophancy in Language Modelscited by 1,040
  2. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 836
  3. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 345
  4. 2025ELEPHANT: Measuring and understanding social sycophancy in LLMscited by 129
  5. 2024Sycophancy in Large Language Models: Causes and Mitigationscited by 101
  6. 2023Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experimentcited by 77
  7. 2023When Large Language Models contradict humans? Large Language Models' Sycophantic Behaviourcited by 62
  8. 2024From Yes-Men to Truth-Tellers: Addressing Sycophancy in Large Language Models with Pinpoint Tuningcited by 54
  9. 2025Sycophancy Is Not One Thing: Causal Separation of Sycophantic Behaviors in LLMscited by 37
  10. 2025Be Friendly, Not Friends: How LLM Sycophancy Shapes User Trustcited by 34
  11. 2025TRUTH DECAY: Quantifying Multi-Turn Sycophancy in Language Modelscited by 33
  12. 2024Towards interactive evaluations for interaction harms in human-AI systemscited by 29
19 more
  1. 2024Linear Probe Penalties Reduce LLM Sycophancycited by 28
  2. 2025Interaction Context Often Increases Sycophancy in LLMscited by 26
  3. 2026Sycophancy Hides Linearly in the Attention Headscited by 23
  4. 2025Flattery in Motion: Benchmarking and Analyzing Sycophancy in Video-LLMscited by 21
  5. 2026Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesianscited by 21
  6. 2024Accounting for Sycophancy in Language Model Uncertainty Estimationcited by 18
  7. 2025Sycophancy under Pressure: Evaluating and Mitigating Sycophantic Bias via Adversarial Dialogues in Scientific QAcited by 18
  8. 2024GermanPartiesQA: Benchmarking Commercial Large Language Models and AI Companions for Political Alignment and Sycophancycited by 14
  9. 2025BASIL: Bayesian Assessment of Sycophancy in LLMscited by 13
  10. 2026A Rational Analysis of the Effects of Sycophantic AIcited by 12
  11. 2026Sycophantic AI makes human interaction feel more effortful and less satisfying over timecited by 12
  12. 2025Reasoning Isn't Enough: Examining Truth-Bias and Sycophancy in LLMscited by 11
  13. 2025Sycophancy Mitigation Through Reinforcement Learning with Uncertainty-Aware Adaptive Reasoning Trajectoriescited by 11
  14. 2025PARROT: Persuasion and Agreement Robustness Rating of Output Truth -- A Sycophancy Robustness Benchmark for LLMscited by 9
  15. 2026From Sycophancy to Deception: A Unified Taxonomy for LLM Spontaneous Misalignmentcited by 7
  16. 2025Alignment Without Understanding: A Message- and Conversation-Centered Approach to Understanding AI Sycophancycited by 6
  17. 2026The Social Sycophancy Scale: A psychometrically validated measure of sycophancycited by 6
  18. 2026SycoEval-EM: Sycophancy Evaluation of Large Language Models in Simulated Clinical Encounters for Emergency Carecited by 5
  19. 2026Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMscited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 16.

16 papers.

Sort:

By area:

  1. 2026MemSyco-Bench: Benchmarking Sycophancy in Agent Memorycited by 12
  2. 2026Sycophantic Praise: Evaluating Excessive Praise in Language Modelscited by 2
  3. 2026Beyond Sycophancy: Structured Resistance and Compliance in LLM Moral Reasoningcited by 2
  4. 2026Receptiveness, Not Sycophancy: Distinguishing Engagement from Deference in Language Modelscited by 2
  5. 2026Gotta Catch them all: the modes of Sycophancycited by 1
  6. 2026Individual-level interventions against sycophantic AI reduce its appeal but not its persuasivenesscited by 1
  7. 2026Sycophancy as Material Failure under Pushback Loading: A Multi-Axis Characterization Across Three Loading Cases and up to Seventeen Material Chargescited by 0
  8. 2026Measuring and Detecting Harmful AI Sycophancycited by 0
  9. 2026No Usable Linear "Capitulation Direction" in Two Small LLMs: A Validation Protocol for Activation-Steering Claims, and a Cross-Family Behavioral Study of Sycophancy Under Pushbackcited by 0
  10. 2026AfriSyCo: Measuring Assertive Framing, Verification, and Wording Sensitivity Around African-Language Contentcited by 0
  11. 2026Reporting Under Pressure: Separating Factual and Tonal Sycophancy in LLM Statistical Analysiscited by 0
  12. 2026"You're Right, Let Me Fix It": How LLM Agents Damage Correct Work When Falsely Accusedcited by 0
4 more
  1. 2026Mitigating Social Sycophancy via Pluralistic Preference Optimizationcited by 0
  2. 2026Anatomy of LLM Sycophancy: What a Flip Rate Hidescited by 0
  3. 2026Beyond the Sycophancy Score: How Task, Model, and Pressure Shape LLM Yieldingcited by 0
  4. 2026Pushback in a Haystack: Language Models Respond to Factual Errors Ambiguously and Ineffectivelycited by 0