Watermarking Language Models for Many Adaptive Users

• Published in: arXiv.org 2024-06
• Main Authors: Cohen, Aloni, Hoover, Alexander, Schoenbach, Gabe
• Format: Article
• Language: English
• Subjects: Black boxes; Robustness; Tracing; Watermarking
• ISSN: 2331-8422

We study watermarking schemes for language models with provable guarantees. As we show, prior works offer no robustness guarantees against adaptive prompting: when a user queries a language model more than once, as even benign users do. And with just a single exception (Christ and Gunn, 2024), prior works are restricted to zero-bit watermarking: machine-generated text can be detected as such, but no additional information can be extracted from the watermark. Unfortunately, merely detecting AI-generated text may not prevent future abuses. We introduce multi-user watermarks, which allow tracing model-generated text to individual users or to groups of colluding users, even in the face of adaptive prompting. We construct multi-user watermarking schemes from undetectable, adaptively robust, zero-bit watermarking schemes (and prove that the undetectable zero-bit scheme of Christ, Gunn, and Zamir (2024) is adaptively robust). Importantly, our scheme provides both zero-bit and multi-user assurances at the same time. It detects shorter snippets just as well as the original scheme, and traces longer excerpts to individuals. The main technical component is a construction of message-embedding watermarks from zero-bit watermarks. Ours is the first generic reduction between watermarking schemes for language models. A challenge for such reductions is the lack of a unified abstraction for robustness -- that marked text is detectable even after edits. We introduce a new unifying abstraction called AEB-robustness. AEB-robustness provides that the watermark is detectable whenever the edited text "approximates enough blocks" of model-generated output.