Evaluating the False Trust engendered by LLM Explanations
For developers and users of LLMs, this work highlights that common explanation types can engender false trust, whereas dual explanations offer a more reliable way to support user decision-making.
The paper investigates whether LLM explanations (reasoning traces, summaries, post-hoc explanations) help users detect incorrect AI answers or merely persuade them to trust the AI. A user study found that reasoning traces and post-hoc explanations increase acceptance regardless of correctness, while dual explanations (arguments for and against) improve users' ability to distinguish correct from incorrect outputs.
Large Language Models (LLMs) and Large Reasoning Models (LRMs) are increasingly used for critical tasks, yet they provide no guarantees about the correctness of their solutions. Users must decide whether to trust the model's answer, aided by reasoning traces, their summaries, or post-hoc generated explanations. These reasoning traces, despite evidence that they are neither faithful representations of the model's computations nor necessarily semantically meaningful, are often interpreted as provenance explanations. It is unclear whether explanations or reasoning traces help users identify when the AI is incorrect, or whether they simply persuade users to trust the AI regardless. In this paper, we take a user-centered approach and develop an evaluation protocol to study how different explanation types affect users' ability to judge the correctness of AI-generated answers and engender false trust in the users. We conduct a between-subject user study, simulating a setting where users do not have the means to verify the solution and analyze the false trust engendered by commonly used LLM explanations - reasoning traces, their summaries and post-hoc explanations. We also test a contrastive dual explanation setting where we present arguments for and against the AI's answer. We find that reasoning traces and post-hoc explanations are persuasive but not informative: they increase user acceptance of LLM predictions regardless of their correctness. In contrast, dual explanation is the only condition that genuinely improves users' ability to distinguish correct from incorrect AI outputs.