TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection

TL;DR

TextSeal is a novel watermarking method for large language models that ensures robust, distortion-free detection of AI-generated text without affecting model performance or inference speed.

Contribution

It introduces a dual-key, localized watermarking scheme that outperforms existing baselines and is robust against dilution and distillation, with no inference overhead.

Findings

TextSeal outperforms SynthID-text in detection strength.

It maintains downstream performance and perceptual quality.

The watermark transfers through model distillation, enabling detection of unauthorized use.

Abstract

We introduce TextSeal, a state-of-the-art watermark for large language models. Building on Gumbel-max sampling, TextSeal introduces dual-key generation to restore output diversity, along with entropy-weighted scoring and multi-region localization for improved detection. It supports serving optimizations such as speculative decoding and multi-token prediction, and does not add any inference overhead. TextSeal strictly dominates baselines like SynthID-text in detection strength and is robust to dilution, maintaining confident localized detection even in heavily mixed human/AI documents. The scheme is theoretically distortion-free, and evaluation across reasoning benchmarks confirms that it preserves downstream performance; while a multilingual human evaluation (6000 A/B comparisons, 5 languages) shows no perceptible quality difference. Beyond its use for provenance detection, TextSeal is also ``radioactive'': its watermark signal transfers through model distillation, enabling detection of unauthorized use.