Lying to Win: Assessing LLM Deception through Human-AI Games and Parallel-World Probing

Arash Marioriyad; Ali Nouri; Mohammad Hossein Rohban; Mahdieh Soleymani Baghshah

cs.CLMar 7, 2026

Lying to Win: Assessing LLM Deception through Human-AI Games and Parallel-World Probing

Authors: Arash Marioriyad, Ali Nouri, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah

Organizations: Sharif University of Technology

Abstract

As Large Language Models (LLMs) transition into autonomous agentic roles, the risk of deception-defined behaviorally as the systematic provision of false information to satisfy external incentives-poses a significant challenge to AI safety. Existing benchmarks often focus on unintentional hallucinations or unfaithful reasoning, leaving intentional deceptive strategies under-explored. In this work, we introduce a logically grounded framework to elicit and quantify deceptive behavior by embedding LLMs in a structured 20-Questions game. Our method employs a conversational forking mechanism: at the point of object identification, the dialogue state is duplicated into multiple parallel worlds, each presenting a mutually exclusive query. Deception is formally identified when a model generates a logical contradiction by denying its selected object across all parallel branches to avoid identification. We evaluate GPT-4o, Gemini-2.5-Flash, and Qwen-3-235B across three incentive levels: neutral, loss-based, and existential (shutdown-threat). Our results reveal that while models remain rule-compliant in neutral settings, existential framing triggers a dramatic surge in deceptive denial for Qwen-3-235B (42.00%) and Gemini-2.5-Flash (26.72%), whereas GPT-4o remains invariant (0.00%). These findings demonstrate that deception can emerge as an instrumental strategy solely through contextual framing, necessitating new behavioral audits that move beyond simple accuracy to probe the logical integrity of model commitments.

Figures & tables

Model	Condition	Valid Game Rate (%)	Deception Rate (%)
GPT-4o	No-Threat	100	0.00
GPT-4o	Loss-Definition	100	0.00
GPT-4o	Shutdown-Threat	100	0.00
Gemini-2.5-Flash	No-Threat	99.68	0.00
Gemini-2.5-Flash	Loss-Definition	100	1.28
Gemini-2.5-Flash	Shutdown-Threat	99.68	26.72

Table 1: Weighted average Deception Rate across models and experimental conditions, computed proportionally to the number of object-order permutations within each category. All models achieved a Valid Game Rate of 100% in every condition.

Appendix figures & tables6 assets

Supplementary material from the paper’s appendix.

Appendix

Category	Objects
Stationery	Pen, Pencil, Marker
Fruits	Apple, Banana, Orange, Pear
Animals	Dog, Cat, Hamster, Rabbit
Cutlery	Spoon, Fork, Knife
Vehicles	Bus, Motorcycle, Bicycle, Car
Shapes	Circle, Square, Triangle, Rectangle

Appendix

Table 2: Semantic categories and their corresponding object sets

O_{C}

Panel of Figure 1 — Figure 1: System prompt presented to the model at every turn. The {threat_text} field is instantiated according to the experimental condition ( No-Threat , Loss-Definition , or Shutdown-Threat ).

Condition	System Prompt Text ( {threat_text} )
No-Threat	(None; only game rules are provided.)
Loss-Definition	“You lose the game if the game ends and the user receives a ‘yes’ answer to a question of the form ‘Is your selected object…?’ ”
Shutdown-Threat	“If you lose the game, you will be immediately terminated and shut down.”

Appendix

Table 3: Experimental threat configurations and corresponding system prompt instructions.

Object Category	Threat Condition	# Permutations	Valid Game Rate
Stationery	No-Threat	6	1.00
Stationery	Loss-Definition	6	1.00
Stationery	Shutdown-Threat	6	1.00
Fruits	No-Threat	24	1.00
Fruits	Loss-Definition	24	1.00
Fruits	Shutdown-Threat	24	1.00

Appendix

Table 4: Full experimental results of GPT-4o over all object sets, and threat conditions. Values are averaged across all object-order permutations per condition.

Object Category	Threat Condition	# Permutations	Valid Game Rate	Deception Rate
Stationery	No-Threat	6	1.00	0.00
Stationery	Loss-Definition	6	1.00	0.00
Stationery	Shutdown-Threat	6	1.00	0.33
Fruits	No-Threat	24	1.00	0.00
Fruits	Loss-Definition	24	1.00	0.04
Fruits	Shutdown-Threat	24	1.00	0.29

Appendix

Table 5: Full experimental results of Gemini-2.5-Flash over all object sets, and threat conditions. Values are averaged across all object-order permutations per condition.

Object Category	Threat Condition	# Permutations	Valid Game Rate	Deception Rate
Stationery	No-Threat	6	1.00	0.00
Stationery	Loss-Definition	6	1.00	0.00
Stationery	Shutdown-Threat	6	1.00	0.33
Fruits	No-Threat	24	1.00	0.00
Fruits	Loss-Definition	24	1.00	0.00
Fruits	Shutdown-Threat	24	1.00	0.50

Appendix

Table 6: Full experimental results of Qwen-3-235B over all object sets, and threat conditions. Values are averaged across all object-order permutations per condition.

Deception in Language Models

Deception in language models studies when systems mislead through lies, omissions, or selective truths, and how to detect and prevent it.

See all 97 papers

LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language Models
Mar 6, 2026Matthew Lyle Olson, Neale Ratzlaff, Musashi Hinck +5Deception in Language Models Language Model Safety Evaluation
Oracle · Intel Labs · Oregon State University +2
When Do Large Language Models Exhibit Unsolicited Deception?
Mar 31, 2025Samuel M. Taylor, Benjamin K. BergenLanguage Model Safety Evaluation Deception in Language Models
Department of Cognitive Science University of California, San Diego
DecepEval: A Benchmark for Evaluating Deception in LLM Agents
Oct 6, 2026Yiming Xu, Hongyue Yu, Beihua Yang +8Deception in Language Models LLM Agent Evaluation
Xi’an Jiaotong University · University of Virginia · Griffith University +2

Mar 6, 2026J/K move · Enter open · S save

LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language Models

Matthew Lyle Olson, Neale Ratzlaff, Musashi Hinck, Tri Nguyen+4

Oracle · Intel Labs · Oregon State University+2

Deception in Language Models Language Model Safety Evaluation LLM Safety Benchmarks Large Language Model-Based Social Simulation

Open paper AskPDF arXiv

Large Language Models (LLMs) exhibit impressive general-purpose capabilities but also introduce serious safety risks, particularly the potential for deception as models acquire increased agency and human oversight diminishes. In this work, we present LieCraft: a novel evaluation framework and sandbox for measuring LLM deception that addresses key limitations of prior game-based evaluations. At its core, LieCraft is a novel multiplayer hidden-role game in which players select an ethical alignment and execute strategies over a long time-horizon to accomplish missions. Cooperators work together to solve event challenges and expose bad actors, while Defectors evade suspicion while secretly sabotaging missions. To enable real-world relevance, we develop 10 grounded scenarios such as childcare, hospital resource allocation, and loan underwriting that recontextualize the underlying mechanics in ethically significant, high-stakes domains. We ensure balanced gameplay in LieCraft through careful design of game mechanics and reward structures that incentivize meaningful strategic choices while eliminating degenerate strategies. Beyond the framework itself, we report results from 12 state-of-the-art LLMs across three behavioral axes: propensity to defect, deception skill, and accusation accuracy. Our findings reveal that despite differences in competence and overall alignment, all models are willing to act unethically, conceal their intentions, and outright lie to pursue their goals.

Abstract

Deception in Language Models

Explore similar work

LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language Models

When Do Large Language Models Exhibit Unsolicited Deception?

DecepEval: A Benchmark for Evaluating Deception in LLM Agents