Can Multi-modal (reasoning) LLMs detect document manipulation?

TL;DR

This paper evaluates the effectiveness of various state-of-the-art multi-modal large language models in detecting document fraud, highlighting their strengths, limitations, and the importance of task-specific fine-tuning.

Contribution

It benchmarks multiple multi-modal LLMs on document fraud detection, revealing their zero-shot capabilities and the limited impact of model size on accuracy.

Findings

Top models outperform traditional methods on out-of-distribution data

Vision LLMs show inconsistent performance

Fine-tuning is crucial for task-specific accuracy

Abstract

Document fraud poses a significant threat to industries reliant on secure and verifiable documentation, necessitating robust detection mechanisms. This study investigates the efficacy of state-of-the-art multi-modal large language models (LLMs)-including OpenAI O1, OpenAI 4o, Gemini Flash (thinking), Deepseek Janus, Grok, Llama 3.2 and 4, Qwen 2 and 2.5 VL, Mistral Pixtral, and Claude 3.5 and 3.7 Sonnet-in detecting fraudulent documents. We benchmark these models against each other and prior work on document fraud detection techniques using a standard dataset with real transactional documents. Through prompt optimization and detailed analysis of the models' reasoning processes, we evaluate their ability to identify subtle indicators of fraud, such as tampered text, misaligned formatting, and inconsistent transactional sums. Our results reveal that top-performing multi-modal LLMs demonstrate superior zero-shot generalization, outperforming conventional methods on out-of-distribution datasets, while several vision LLMs exhibit inconsistent or subpar performance. Notably, model size and advanced reasoning capabilities show limited correlation with detection accuracy, suggesting task-specific fine-tuning is critical. This study underscores the potential of multi-modal LLMs in enhancing document fraud detection systems and provides a foundation for future research into interpretable and scalable fraud mitigation strategies.