Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers

ACL ARR 2025 May Submission827 Authors

Abstract: We study privacy leakage in the reasoning traces of large reasoning models used as personal agents which handle sensitive user data. Unlike final outputs, reasoning traces are often assumed to be internal and safe. We challenge this assumption by showing that reasoning traces frequently contain sensitive user data, which can be extracted via prompt injections or accidentally leak into outputs. Through probing and agentic evaluations, we demonstrate that test-time compute approaches, particularly increased reasoning steps, amplify such leakage. While increasing the budget of those test-time compute approaches makes models more cautious in their final answers, it also leads them to reason more verbosely and leak more in their own thinking. This reveals a core tension: reasoning improves utility but enlarges the privacy attack surface. We argue that safety efforts must extend to the model's internal thinking, not just its outputs.

Paper Type: Long

Research Area: Language Modeling

Research Area Keywords: security and privacy,chain-of-thought,scaling,LLM/AI agents,safety and alignment

Contribution Types: Model analysis & interpretability, NLP engineering experiment, Reproduction study, Publicly available software and/or pre-trained models, Data resources, Data analysis

Languages Studied: English

Keywords: security and privacy, chain-of-thought, scaling, LLM/AI agents, safety and alignment

Submission Number: 827