Data Provenance for Image Auto-Regressive Generation

2026-06-22

AI Summary

This work addresses the lack of reliable provenance mechanisms for image autoregressive (IAR) generative models by proposing a post-hoc attribution framework that requires neither modifications to the generation pipeline nor embedded watermarks. The method leverages intrinsic autoregressive patterns present in IAR-generated images to identify their source model after publication. Without relying on model alterations or auxiliary annotations, the framework demonstrates high accuracy and strong robustness across multiple state-of-the-art IAR architectures. To the best of our knowledge, this is the first approach capable of effectively tracing existing IAR-generated content, offering a practical technical pathway for combating disinformation and enabling accountability for harmful synthetic media.

Abstract

Image autoregressive models (IARs) have recently demonstrated remarkable capabilities in visual content generation, achieving photorealistic quality and rapid synthesis through the next-token prediction paradigm adapted from large language models. As these models become widely accessible, robust data provenance is required to reliably trace IAR-generated images to the source model that synthesized them. This is critical to prevent the spread of misinformation, detect fraud, and attribute harmful content. We find that although IAR-generated images often appear visually identical to real images, their generation process introduces characteristic patterns in their outputs, which serves as a reliable provenance signal for the generated images. Leveraging this, we present a post-hoc framework that enables the robust detection of such patterns for provenance tracing. Notably, our framework does not require modifications of the generative process or outputs. Thereby, it is applicable in contexts where prior watermarking methods cannot be used, such as for generated content that is already published without additional marks and for models that do not integrate watermarking. We demonstrate the effectiveness of our approach across a wide range of IARs, highlighting its high potential for robust data provenance tracing in autoregressive image generation.

Problem

• data provenance
• image autoregressive generation
• source tracing
• misinformation detection
• AI-generated images

Innovation

• data provenance
• image autoregressive models
• post-hoc detection
• generation fingerprinting
• watermark-free tracing