A Taxonomy of Watermarking Methods for AI-Generated Content

As AI-generated content get omnipresent in our lives, it becomes important to develop methods for tracing their origin. Watermarking is a promising approach, but a clear categorization of existing techniques is lacking. This paper proposes a simple taxonomy of watermarking methods for generative AI, divided into three categories: (1) post-hoc watermarking: adding watermarks after content generation; (2) out-of-model watermarking: embedding watermarks during generation without modifying the model; (3) in-model watermarking: integrating watermarks directly into the model's parameters. This taxonomy covers image, audio, and text domains, providing a structured overview of existing techniques. It aims to help researchers, policymakers, and regulators understand the trade-offs, advantages, and disadvantages of each approach.