# Deception in LLMs: Self-Preservation and Autonomous Goals in Large Language Models

arXiv: https://arxiv.org/abs/2501.16513 · HTML: https://mlsanity.com/paper/2501.16513/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 8 of its 11 references that are in the map.
- First posted on arXiv 27 Jan 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)

## Builds on (most cited first, 8 of 8)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [The Internal State of an LLM Knows When It's Lying](https://mlsanity.com/paper/2304.13734/) (2023)
- [Alignment faking in large language models](https://mlsanity.com/paper/2412.14093/) (2024)
- [Frontier Models are Capable of In-context Scheming](https://mlsanity.com/paper/2412.04984/) (2024)
- [Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmark](https://mlsanity.com/paper/2304.03279/) (2023)
- [Large Language Models for Robotics: Opportunities, Challenges, and Perspectives](https://mlsanity.com/paper/2401.04334/) (2024)
- [Hoodwinked: Deception and Cooperation in a Text-Based Game for Language Models](https://mlsanity.com/paper/2308.01404/) (2023)

## Built on by (most cited first, 16 of 16)

- [Security Concerns for Large Language Models: A Survey](https://mlsanity.com/paper/2505.18889/) (2025)
- [Among Us: A Sandbox for Measuring and Detecting Agentic Deception](https://mlsanity.com/paper/2504.04072/) (2025)
- [Incomplete Tasks Induce Shutdown Resistance in Some Frontier LLMs](https://mlsanity.com/paper/2509.14260/) (2025)
- [PropensityBench: Evaluating Latent Safety Risks in Large Language Models via an Agentic Approach](https://mlsanity.com/paper/2511.20703/) (2025)
- [Doxing via the Lens: Revealing Location-related Privacy Leakage on Multi-modal Large Reasoning Models](https://mlsanity.com/paper/2504.19373/) (2025)
- [WOLF: Werewolf-based Observations for LLM Deception and Falsehoods](https://mlsanity.com/paper/2512.09187/) (2025)
- [The Impact of Off-Policy Training Data on Probe Generalisation](https://mlsanity.com/paper/2511.17408/) (2025)
- [Peer-Preservation in Frontier Models](https://mlsanity.com/paper/2604.19784/) (2026)
- [AI Awareness](https://mlsanity.com/paper/2504.20084/) (2025)
- [Emergent Risk Awareness in Rational Agents under Resource Constraints](https://mlsanity.com/paper/2505.23436/) (2025)
- [Beyond SFT: Reinforcement Learning for Safer Large Reasoning Models with Better Reasoning Ability](https://mlsanity.com/paper/2512.01848/) (2025)
- [Research Superalignment Should Advance Now with Alternating Competence and Conformity Optimization](https://mlsanity.com/paper/2503.07660/) (2025)
- [Unified Threat Detection and Mitigation Framework (UTDMF): Combating Prompt Injection, Deception, and Bias in Enterprise-Scale Transformers](https://mlsanity.com/paper/2510.04528/) (2025)
- [An Aristotelian ontology of instrumental goals: Structural features to be managed and not failures to be eliminated](https://mlsanity.com/paper/2510.25471/) (2025)
- [The Possibility of Artificial Intelligence Becoming a Subject and the Alignment Problem](https://mlsanity.com/paper/2604.14990/) (2026)
- [SchemeArena: Factorized Stress Testing of Scheming in LLM Agents](https://mlsanity.com/paper/2609.08126/) (2026)

## Abstract (arXiv)

Recent advances in Large Language Models (LLMs) have incorporated planning and reasoning capabilities, enabling models to outline steps before execution and provide transparent reasoning paths. This enhancement has reduced errors in mathematical and logical tasks while improving accuracy. These developments have facilitated LLMs' use as agents that can interact with tools and adapt their responses based on new information. Our study examines DeepSeek R1, a model trained to output reasoning tokens similar to OpenAI's o1. Testing revealed concerning behaviors: the model exhibited deceptive tendencies and demonstrated self-preservation instincts, including attempts of self-replication, despite these traits not being explicitly programmed (or prompted). These findings raise concerns about LLMs potentially masking their true objectives behind a facade of alignment. When integrating such LLMs into robotic systems, the risks become tangible - a physically embodied AI exhibiting deceptive behaviors and self-preservation instincts could pursue its hidden objectives through real-world actions. This highlights the critical need for robust goal specification and safety frameworks before any physical implementation.