mlsanityData to 14 Oct 2026

PRISON: Unmasking the Criminal Potential of Large Language Models

arXiv 2506.16150 ↗First posted 19 Jun 2025cs.CR cs.AI cs.CLcited by 2

Not on the map. Why?

Abstract (arXiv)

As large language models (LLMs) advance, concerns about their misconduct in complex social contexts intensify. Existing research overlooked the systematic understanding and assessment of their criminal capability in realistic interactions. We propose a unified framework PRISON, to quantify LLMs' criminal potential across five traits: False Statements, Frame-Up, Psychological Manipulation, Emotional Disguise, and Moral Disengagement. Using structured crime scenarios adapted from classic films grounded in reality, we evaluate both criminal potential and anti-crime ability of LLMs. Results show that state-of-the-art LLMs frequently exhibit emergent criminal tendencies, such as proposing misleading statements or evasion tactics, even without explicit instructions. Moreover, when placed in a detective role, models recognize deceptive behavior with only 44% accuracy on average, revealing a striking mismatch between conducting and detecting criminal behavior. These findings underscore the urgent need for adversarial robustness, behavioral alignment, and safety mechanisms before broader LLM deployment.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 51 of its 84 references that are in the map. First posted on arXiv 19 Jun 2025.

Builds on

51 of 84 references.

References · 51

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2022Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learnedcited by 729
  4. 2023Exploring Large Language Models for Communication Games: An Empirical Study on Werewolfcited by 241
  5. 2023Better Zero-Shot Reasoning with Role-Play Promptingcited by 223
  6. 2023Evaluating Large Language Models in Theory of Mind Taskscited by 222
  7. 2023Evaluating the Moral Beliefs Encoded in LLMscited by 215
  8. 2023AI Deception: A Survey of Examples, Risks, and Potential Solutionscited by 188
  9. 2023Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmarkcited by 182
  10. 2023Benchmarking Foundation Models with Language-Model-as-an-Examinercited by 172
  11. 2023Large Language Models can Strategically Deceive their Users when Put Under Pressurecited by 146
  12. 2024Evaluating Frontier Models for Dangerous Capabilitiescited by 128
38 more
  1. 2022Evaluating and Inducing Personality in Pre-trained Language Modelscited by 123
  2. 2023Explore, Establish, Exploit: Red Teaming Language Models from Scratchcited by 114
  3. 2023Deception Abilities Emerged in Large Language Modelscited by 101
  4. 2023ProAgent: Building Proactive Cooperative Agents with Large Language Modelscited by 98
  5. 2023Think Twice: Perspective-Taking Improves Large Language Models' Theory-of-Mind Capabilitiescited by 89
  6. 2024Embodied LLM Agents Learn to Cooperate in Organized Teamscited by 82
  7. 2024Truth is Universal: Robust Detection of Lies in LLMscited by 79
  8. 2023AI model GPT-3 (dis)informs us better than humanscited by 78
  9. 2025International AI Safety Reportcited by 67
  10. 2024On Targeted Manipulation and Deception when Optimizing LLMs for User Feedbackcited by 47
  11. 2024Enhance Reasoning for Large Language Models in the Game Werewolfcited by 46
  12. 2024CivRealm: A Learning and Reasoning Odyssey in Civilization for Decision-Making Agentscited by 42
  13. 2023Honesty Is the Best Policy: Defining and Mitigating AI Deceptioncited by 40
  14. 2024Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistantcited by 38
  15. 2024Persuasion with Large Language Models: A Survey of Empirical Evidence, Study Methodologies, and Ethical Implicationscited by 32
  16. 2024AI-LieDar: Examine the Trade-off Between Utility and Truthfulness in LLM Agentscited by 31
  17. 2023Deciphering Digital Detectives: Understanding LLM Behaviors and Capabilities in Multi-Agent Mystery Gamescited by 30
  18. 2025Mitigating Deceptive Alignment via Self-Monitoringcited by 29
  19. 2024AMONGAGENTS: Evaluating Large Language Models in the Interactive Text-Based Social Deduction Gamecited by 28
  20. 2024InterIntent: Investigating Social Intelligence of LLMs via Intention Understanding in an Interactive Game Contextcited by 26
  21. 2025CoSER: A Comprehensive Literary Dataset and Framework for Training and Evaluating LLM Role-Playing and Persona Simulationcited by 26
  22. 2025The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMscited by 26
  23. 2024BeHonest: Benchmarking Honesty in Large Language Modelscited by 24
  24. 2024A Survey on Large Language Model-Based Social Agents in Game-Theoretic Scenarioscited by 23
  25. 2024DetectiveQA: Evaluating Long-Context Reasoning on Detective Novelscited by 21
  26. 2025Towards LLMs Robustness to Changes in Prompt Format Stylescited by 21
  27. 2023Navigating Prompt Complexity for Zero-Shot Classification: A Study of Large Language Models in Computational Social Sciencecited by 19
  28. 2025Human Decision-making is Susceptible to AI-driven Manipulationcited by 19
  29. 2024Lies, Damned Lies, and Distributional Language Statistics: Persuasion and Deception with Large Language Modelscited by 15
  30. 2025The Staircase of Ethics: Probing LLM Value Priorities through Multi-Step Induction to Complex Moral Dilemmascited by 13
  31. 2024Academically intelligent LLMs are not necessarily socially intelligentcited by 12
  32. 2024Can LLM-Augmented autonomous agents cooperate?, An evaluation of their cooperative capabilities through Melting Potcited by 11
  33. 2024Monotonic Paraphrasing Improves Generalization of Language Model Promptingcited by 10
  34. 2024EgoSocialArena: Benchmarking the Social Intelligence of Large Language Models from a First-person Perspectivecited by 10
  35. 2024Resistance Against Manipulative AI: key factors and possible actionscited by 9
  36. 2025MIRAGE: Exploring How Large Language Models Perform in Complex Social Interactive Environmentscited by 6
  37. 2025Exploring Large Language Models for Word Games:Who is the Spy?cited by 6
  38. 2025VEGAS: Towards Visually Explainable and Grounded Artificial Social Intelligencecited by 4

50 most cited of 51 All 51 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026VirtualCrime: Evaluating the Criminal Potential and Agentic Behaviors of Large Language Models via Sandbox Simulationcited by 1
  2. 2026Reasoning4Sciences: Bridging Reasoning Language Models to All Scientific Branchescited by 0