Can Generative Models Actually Forge Realistic Identity Documents?
This addresses concerns about misuse of generative AI for document forgery, providing an incremental assessment of realistic risks for security and verification stakeholders.
This paper investigates whether current open-source diffusion models can create realistic identity document forgeries that could bypass verification systems, finding that while they mimic surface aesthetics, they fail to achieve structural and forensic authenticity, suggesting the risk may be overestimated.
Generative image models have recently shown significant progress in image realism, leading to public concerns about their potential misuse for document forgery. This paper explores whether contemporary open-source and publicly accessible diffusion-based generative models can produce identity document forgeries that could realistically bypass human or automated verification systems. We evaluate text-to-image and image-to-image generation pipelines using multiple publicly available generative model families, including Stable Diffusion, Qwen, Flux, Nano-Banana, and others. The findings indicate that while current generative models can simulate surface-level document aesthetics, they fail to reproduce structural and forensic authenticity. Consequently, the risk of generative identity document deepfakes achieving forensic-level authenticity may be overestimated, underscoring the value of collaboration between machine learning practitioners and document-forensics experts in realistic risk assessment.