19 August 2026

Claude watermark removed within hours of rollout

First reported

Prompt Engineering Daily and Wired ran this on 19 August, all on the same day.

Anthropic added invisible watermarks to Claude text to comply with EU rules requiring AI-generated content be machine-detectable, with fines up to 3% of annual revenue for non-compliance.
Developer Guillaume Meyer published code removing the watermarks within four hours, gaining 20,000 bookmarks on X and over 100 contributors adapting it for their own projects.
Meyer and others object to invisible watermarking, citing risks of false positives affecting job applicants and researchers, while some simply enjoy the technical challenge of defeating it.
Multiple independent removal methods already exist using techniques like synonym replacement, sentence reordering, and translation between languages, suggesting the watermarks may not be durable long-term.

How it was covered

A Lloyds Business Barometer survey found 54% of UK employers say AI has created jobs at their company, with a quarter now hiring for AI skills and a fifth creating entirely new AI-related roles. However, larger firms with over £10 million turnover are better positioned to benefit, with two-thirds saying their current staff already have needed AI skills, suggesting uneven job creation across company sizes.

Reported by Wired