DeceptionX: Explainable Deception Detection with Multimodal Large Language Models

2026-06-09

AI Summary

This work addresses the limited interpretability of existing deception detection methods, which typically rely on black-box models and fail to emulate human experts’ logical reasoning based on multimodal cues. To bridge this gap, the authors propose DeceptionX, a novel framework that reframes deception detection as an interpretable “observe–reason–conclude” inference process. Central to this approach is DeceptChain, a high-quality dataset of human feedback, which enables the integration of fine-grained audiovisual evidence with chain-of-thought reasoning. The method further incorporates a three-stage training scheme and a Difference-Aware Redundancy Elimination (DARE) strategy to enhance generalization. Experiments demonstrate that DeceptionX outperforms current multimodal large language model (MLLM) baselines and state-of-the-art methods on real-world benchmarks, while simultaneously generating transparent, expert-level reasoning traces—thus achieving a unified balance between high accuracy and strong interpretability.

Abstract

Deception detection is a critical and highly challenging task within affective computing and behavioral analysis. Existing deep learning methods typically treat this task as a straightforward classification problem; however, this black-box approach lacks interpretability and fails to capture the complex logical deduction processes utilized by human experts when identifying lies. While Multimodal Large Language Models (MLLMs) have shown potential, applying them effectively requires a bridge between low-level audiovisual cues and high-level logical reasoning. In this paper, we propose DeceptionX, a novel MLLM framework that shifts the paradigm of deception detection from black-box classification to an interpretable Observe-Think-Summarize reasoning process. To address the scarcity of high-quality reasoning data, we first constructed DeceptChain, a high-quality dataset developed through a human-in-the-loop process. This dataset synthesizes fine-grained visual and auditory evidence (such as micro-expressions and vocal tremors) into structured chain-of-thought reasoning data. Furthermore, we propose a three-stage training pipeline and a Discrepancy-Aware Redundancy Elimination~(DARE) strategy for DeceptionX to further enhance the model's generalization capabilities. Extensive experiments demonstrate that DeceptionX not only outperforms existing MLLM baselines and state-of-the-art methods on standard real-world benchmarks but also provides transparent, expert-level reasoning paths, bridging the critical gap between accuracy and interpretability in multimodal deception detection.

Problem

• deception detection
• interpretability
• multimodal reasoning
• black-box classification
• affective computing

Innovation

• Multimodal Large Language Models
• Explainable AI
• Chain-of-Thought Reasoning
• Deception Detection
• Human-in-the-Loop Dataset