Towards A Correct Usage of Cryptography in Semantic Watermarks for Diffusion Models
This work clarifies cryptographic ambiguities in semantic watermarking for diffusion models, which is incremental but important for improving security and reliability in AI-generated content.
The paper addresses issues in the cryptographic usage of Gaussian Shading for semantic watermarking in diffusion models, introducing a novel proof of lossless performance based on IND$-CPA security and discussing configurations for security, efficiency, and quality.
Semantic watermarking methods enable the direct integration of watermarks into the generation process of latent diffusion models by only modifying the initial latent noise. One line of approaches building on Gaussian Shading relies on cryptographic primitives to steer the sampling process of the latent noise. However, we identify several issues in the usage of cryptographic techniques in Gaussian Shading, particularly in its proof of lossless performance and key management, causing ambiguity in follow-up works, too. In this work, we therefore revisit the cryptographic primitives for semantic watermarking. We introduce a novel, general proof of lossless performance based on IND\$-CPA security for semantic watermarks. We then discuss the configuration of the cryptographic primitives in semantic watermarks with respect to security, efficiency, and generation quality.