On Google's SynthID-Text LLM Watermarking System: Theoretical Analysis and Empirical Validation

TL;DR

This paper provides a comprehensive theoretical and empirical analysis of Google's SynthID-Text watermarking system, highlighting its detection capabilities, robustness, vulnerabilities, and potential for future improvements in AI-generated text identification.

Contribution

It offers the first theoretical analysis of SynthID-Text, validating its detection performance and robustness, and introduces new insights into watermark removal strategies and system vulnerabilities.

Findings

Mean score is vulnerable to increased tournament layers.

Bayesian score improves watermark robustness.

Optimal detection parameter is set to 0.5.

Abstract

Google's SynthID-Text, the first ever production-ready generative watermark system for large language model, designs a novel Tournament-based method that achieves the state-of-the-art detectability for identifying AI-generated texts. The system's innovation lies in: 1) a new Tournament sampling algorithm for watermarking embedding, 2) a detection strategy based on the introduced score function (e.g., Bayesian or mean score), and 3) a unified design that supports both distortionary and non-distortionary watermarking methods. This paper presents the first theoretical analysis of SynthID-Text, with a focus on its detection performance and watermark robustness, complemented by empirical validation. For example, we prove that the mean score is inherently vulnerable to increased tournament layers, and design a layer inflation attack to break SynthID-Text. We also prove the Bayesian score offers improved watermark robustness w.r.t. layers and further establish that the optimal Bernoulli distribution for watermark detection is achieved when the parameter is set to 0.5. Together, these theoretical and empirical insights not only deepen our understanding of SynthID-Text, but also open new avenues for analyzing effective watermark removal strategies and designing robust watermarking techniques. Source code is available at https: //github.com/romidi80/Synth-ID-Empirical-Analysis.