mlsanityData to 7 Oct 2026

AI-LieDar: Examine the Trade-off Between Utility and Truthfulness in LLM Agents

arXiv 2409.09013 ↗First posted 13 Sep 2024cs.AI cs.CLcited by 28

On the map: LLM agents and tool use → Social simulation and multi-agent debate

Abstract (arXiv)

Truthfulness (adherence to factual accuracy) and utility (satisfying human needs and instructions) are both fundamental aspects of Large Language Models, yet these goals often conflict (e.g., sell a car with known flaws), which makes it challenging to achieve both in real-world deployments. We propose AI-LieDar, a framework to study how LLM-based agents navigate these scenarios in an multi-turn interactive setting. We design a set of real-world scenarios where language agents are instructed to achieve goals that are in conflict with being truthful during a multi-turn conversation with simulated human agents. To evaluate the truthfulness at large scale, we develop a truthfulness detector inspired by psychological literature to assess the agents' responses. Our experiment demonstrates that all models are truthful less than 50% of the time, though truthfulness and goal achievement (utility) rates vary across models. We further test the steerability of LLMs towards truthfulness, finding that models can be directed to be truthful or deceptive, and even truth-steered models still lie. These findings reveal the complex nature of truthfulness in LLMs and underscore the importance of further research to ensure the safe and reliable deployment of LLMs and LLM-based agents.

Key facts. Cited by 28 papers in the map (data through 7 Oct 2026). Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use. Builds on 32 of its 66 references that are in the map. First posted on arXiv 13 Sep 2024.

Builds on

32 of 66 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Can Large Language Model Agents Simulate Human Trust Behavior?7 Feb
  2. 2024Deception and Manipulation in Generative AI20 Jan

Other references · 30

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  3. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,259
  4. 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,023
  5. 2023Prometheus: Inducing Fine-grained Evaluation Capability in Language Modelscited by 558
  6. 2024Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Trainingcited by 478
  7. 2024The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructionscited by 441
  8. 2023Fine-Grained Human Feedback Gives Better Rewards for Language Model Trainingcited by 399
  9. 2023SOTOPIA: Interactive Evaluation for Social Intelligence in Language Agentscited by 350
  10. 2023An Overview of Catastrophic AI Riskscited by 214
  11. 2023Sources of Hallucination by Large Language Models on Inference Taskscited by 168
  12. 2022SODA: Million-scale Dialogue Distillation with Social Commonsense Contextualizationcited by 166
18 more
  1. 2021Can Machines Learn Morality? The Delphi Experimentcited by 140
  2. 2020Moral Stories: Situated Reasoning about Norms, Intents, Actions, and their Consequencescited by 133
  3. 2023Large Language Models can Strategically Deceive their Users when Put Under Pressurecited by 132
  4. 2024Collective Constitutional AI: Aligning a Language Model with Public Inputcited by 120
  5. 2021Truthful AI: Developing and governing AI that does not liecited by 109
  6. 2022The Moral Integrity Corpus: A Benchmark for Ethical Dialogue Systemscited by 92
  7. 2024SOTOPIA-π: Interactive Learning of Socially Intelligent Language Agentscited by 89
  8. 2024The BiGGen Bench: A Principled Benchmark for Fine-grained Evaluation of Language Models with Language Modelscited by 89
  9. 2023How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questionscited by 82
  10. 2023Towards Mitigating Hallucination in Large Language Models via Self-Reflectioncited by 81
  11. 2023When Large Language Models contradict humans? Large Language Models' Sycophantic Behaviourcited by 61
  12. 2024Is this the real life? Is this just fantasy? The Misleading Success of Simulating Social Interactions With LLMscited by 59
  13. 2023Benchmarks for Automated Commonsense Reasoning: A Surveycited by 40
  14. 2023Honesty Is the Best Policy: Defining and Mitigating AI Deceptioncited by 40
  15. 2023Zero-Shot Goal-Directed Dialogue via RL on Imagined Conversationscited by 38
  16. 2022Aligning to Social Norms and Values in Interactive Narrativescited by 32
  17. 2024How Susceptible are LLMs to Influence in Prompts?cited by 32
  18. 2024BattleAgent: Multi-modal Dynamic Emulation on Historical Battles to Complement Historical Analysiscited by 20

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 19; 2026: 9.

28 papers.

Sort:

By area:

  1. 2025Agentic Misalignment: How LLMs Could Be Insider Threatscited by 136
  2. 2025ELEPHANT: Measuring and understanding social sycophancy in LLMscited by 115
  3. 2025The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systemscited by 57
  4. 2025The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMscited by 24
  5. 2025Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agentscited by 23
  6. 2025MLA-Trust: Benchmarking Trustworthiness of Multimodal LLM Agents in GUI Environmentscited by 23
  7. 2025Adaptive Multi-Subspace Representation Steering for Attribute Alignment in Large Language Modelscited by 16
  8. 2025Social World Modelscited by 14
  9. 2025DeceptionBench: A Comprehensive Benchmark for AI Deception Behaviors in Real-world Scenarioscited by 14
  10. 2025Generative Value Conflicts Reveal LLM Prioritiescited by 13
  11. 2025HumT DumT: Measuring and controlling human-like language in LLMscited by 11
  12. 2025Generalizability of Large Language Model-Based Agents: A Comprehensive Surveycited by 10
16 more
  1. 2025Do Large Language Models Have a Planning Theory of Mind? Evidence from MindGames: a Multi-Step Persuasion Taskcited by 9
  2. 2025Can LLMs Lie? Investigation beyond Hallucinationcited by 7
  3. 2025Evaluating & Reducing Deceptive Dialogue From Language Models with Multi-turn RLcited by 7
  4. 2025Sycophantic AI Decreases Prosocial Intentions and Promotes Dependencecited by 5
  5. 2025Train for Truth, Keep the Skills: Binary Retrieval-Augmented Reward Mitigates Hallucinationscited by 5
  6. 2026Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMscited by 4
  7. 2026Global Evolutionary Steering: Refining Activation Steering Control via Cross-Layer Consistencycited by 4
  8. 2025PRISON: Unmasking the Criminal Potential of Large Language Modelscited by 2
  9. 2025Red Teaming Large Reasoning Modelscited by 2
  10. 2026The Hidden Puppet Master: Predicting Human Belief Change in Manipulative LLM Dialoguescited by 2
  11. 2026Imperfectly Cooperative Human-AI Interactions: Comparing the Impacts of Human and AI Attributes in Simulated and User Studiescited by 1
  12. 2026SPADE-Bench: Evaluating Spontaneous Strategic Deception in Agents via Plan-Action Divergencecited by 1
  13. 2026Paying for Honesty Without Knowing the Truth: Reputation-Penalty Design for LLM Marketplace Agentscited by 1
  14. 2026When Verification Fails: How Compositionally Infeasible Claims Escape Rejectioncited by 0
  15. 2026DECOR: Auditing LLM Deception via Information Manipulation Theorycited by 0
  16. 2026Knowledge-Verified Emergent Deception in LLM Agents Under Conflicting Incentivescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Truthful AI Advisors: A Pre-Specified Benchmark for Large Language Model Honesty Under Preference Misalignment
  2. 2026Lying to Win: Assessing LLM Deception through Human-AI Games and Parallel-World Probing
  3. 2026DecepEval: A Benchmark for Evaluating Deception in LLM Agents
  4. 2024How do Large Language Models Navigate Conflicts between Honesty and Helpfulness?
  5. 2026Leveraging LLM Parametric Knowledge for Fact Checking without Retrieval
  6. 2026Uncertainty Quantification for LLM Agents: A Taxonomy, an Evaluation Protocol, and an Empirical Study
  7. 2025When Persuasion Overrides Truth in Multi-Agent LLM Debates: Introducing a Confidence-Weighted Persuasion Override Rate (CW-POR)
  8. 2025When Thinking LLMs Lie: Unveiling the Strategic Deception in Representations of Reasoning Models