作者位次 #4/5ACL 2026Findings· Findings of ACL
Position: LLM Watermarking Should Align Stakeholders’ Incentives for Practical Adoption
Yepeng Liu · Xuandong Zhao · Dawn Song · Gregory W. Wornell · Yuheng Bu
摘要
Despite progress in watermarking algorithms for large language models (LLMs), real-world deployment remains limited. We argue that this gap stems from misaligned incentives among LLM providers, platforms, and end users, which manifest as three key barriers: competitive risk, detection-tool govern...