Why the Claude AI Watermark Misses the Real Problem
18 Aug 2026
Putting a stigma on progress has never been a good idea. Since the 2nd of August, text written by Claude has an invisible watermark.
Putting a stigma on progress has never been a good idea. Since the 2nd of August, text written by Claude has an invisible watermark.
The timing matches the EU AI Act, which became applicable that same day, and Anthropic applied it worldwide, not only in Europe.
I spend my days helping companies and leaders get AI adoption moving. The blocker is almost never the tool. It is the fear of being caught using it.
A recent study found the same thing. People are more likely to use AI when they feel safe at work.
A hidden mark on everything they write does the opposite.
Let me ask the plain question: what is wrong if the text was generated by Claude?
This is how I see it.
→ Boring or difficult writing was bad writing long before AI
→ I check whether it is useful to the reader. Nothing else.
→ A wrong claim was wrong whoever typed it
AI did not change any of that.
What do you make of the watermark?
🧩 The study I refer to is here https://arxiv.org/abs/2602.23279
🧩 I write about these topics in my bi-weekly newsletter, sign up now to get it on Friday: https://www.thechangerepublic.com/#subscribe
What actually builds psychological safety around AI at work
The pattern I see in almost every organisation I work with is the same. Leadership announces that AI is approved, sometimes even mandated, and assumes that settles the question. It does not. What settles it is whether people believe they can say, out loud, in a team meeting, that they used AI to draft something, without it counting against them later.
A watermark does not fix that. The leaders whose teams actually adopt AI are the ones who go first: naming what they used it for, admitting what it got wrong, showing that using it well is a skill, not a shortcut to hide.
I built my TEDx talk, What AI Can't Hear, around exactly this gap between what AI can do and what only a human presence in the room can do. Watermarks and detection tools try to solve this with technology. What actually moves teams is a leadership habit: making it safe to say ‘I used AI for this’.
If this resonates and you want a practical guide, I have a free What AI Can't Hear Pack that distils seven listening behaviours from my TEDx talk plus ten practical questions to use with your team to talk about AI without fear or confusion. Leave your email here and I will send it over: https://www.thechangerepublic.com/free-resources