Stable Signature is Unstable: Removing Image Watermark from Diffusion Models

TL;DR

This paper demonstrates that the Stable Signature watermarking method for diffusion models can be effectively removed through fine-tuning, challenging its robustness and raising concerns about watermark security.

Contribution

We introduce a fine-tuning attack that successfully removes Stable Signature watermarks from diffusion models without degrading image quality.

Findings

Watermarks can be effectively removed via fine-tuning.

Stable Signature's robustness is less than previously claimed.

Watermarked images can be indistinguishable from non-watermarked ones.

Abstract

Watermark has been widely deployed by industry to detect AI-generated images. A recent watermarking framework called \emph{Stable Signature} (proposed by Meta) roots watermark into the parameters of a diffusion model's decoder such that its generated images are inherently watermarked. Stable Signature makes it possible to watermark images generated by \emph{open-source} diffusion models and was claimed to be robust against removal attacks. In this work, we propose a new attack to remove the watermark from a diffusion model by fine-tuning it. Our results show that our attack can effectively remove the watermark from a diffusion model such that its generated images are non-watermarked, while maintaining the visual quality of the generated images. Our results highlight that Stable Signature is not as stable as previously thought.