Position: LLM Watermarking Should Align Stakeholders’ Incentives for Practical Adoption
arXiv:2510.18333v2 Announce Type: replace-cross
Abstract: 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,…