The AI Watermark Freakout
If your social media feeds look anything like mine, you’ve probably seen a ton of posts from people losing their minds about watermarks and IP and AI slop.
Let's get into it.
Anthropic, the company that created Claude (the AI I use every day, openly, because I’m not weird about it) now leaves an invisible pattern on the text that goes through its little invisible robot fingers. So does Google’s AI. It has since 2024. So does OpenAI (ChatGPT) on its images and probbaly soon also text.
Here's the link to a short video where I describe this a bit.
This is not some scandal. It is fully based on EU regulation that went into effect early last month. And every major AI signed the agreement. And honestly, this is a good thing.
But the people who have been pretending they don’t use AI are spiraling.
And the ones who want you to think that 100% human is the only way to create great stuff are using this as yet another way to bash AI.
But there is a huge misunderstanding about what this watermark is. First of all, it is generated on the words that the AI chooses between (that's explained in the token choice video you can watch in the link above).
If you wrote something yourself and asked AI to clean it up, tighten a paragraph, or fix your grammar… there probably isn’t enough of a pattern to detect. Anthropic said this themselves.
(They also said that EVERYTHING would be marked, and while I can't figure out HOW they would do that if it is using YOUR word choices, I tend to believe them... and that is a much bigger issue around ownership that I'll be keeping an eye on).
Where it really makes a difference, is wherever AI wrote something from scratch. Which, if you use AI the way I teach, would literally never happen.
Let's assume your process is: think your own thoughts, say them out loud, use the transcript in AI to help you shape your words into the type of content you want with the desired results you want (and I'm not just talking about copy or creative output here)...
You’re off to a good start. But here's where my method will help you be not just fine, but excellent.
THE EDIT:
I don't just give AI a transcript when I start working with it. I read the work out loud before I ship it off. And that's where you can literally eradicate this watermark. When your mouth gives your body the signal that a word is not a word you would use, change it. If you do this diligently, the pattern created by this watermark will disappear AND your voice will become stronger.
Yeah, it takes a little extra time, but it's the difference between content that was thought of by you and sounds like everyone else and content that makes you money because it is unique, sharp, and speaking in a way no one else is.
You know who’s NOT fine with all this watermark stuff?
The people who have been copying and pasting AI output directly into their content, their client work, their “thought leadership” 🙃 and pretending they wrote it. And I know a lot of these people who have been charging premium rates for work they outsourced to a chatbot without saying a word about it.
THAT is who this bothers. Good. It should. Let me be clear about something. Using AI is not the problem.
Lying about using AI is the problem.
Using AI poorly is the problem.
Not involving yourself enough in the AI process is the problem.
And the reason so many people are panicking right now is not really because the technology changed, but because their cover just got pulled off.
Meanwhile...
The people shaming others for using AI are often the same people who:
📌 Use Grammarly (AI)
📌 Use spell check (AI)
📌 Use Google’s predictive text in their emails (AI)
📌 Use templates someone else wrote and changed three words (not AI, but just lazy, and nobody’s watermarking that)
The moral high ground here is imaginary.
The watermark does not mean “this is bad content.” It means “AI was probably involved.” Those are two completely different statements. And the people conflating them are basically showing us how little they truly understand about the technology.
Here’s what I think is happening. For years, people built entire identities around being anti-AI. They said it would make writers and creatives obsolete. They said it was cheating. They said it was the end of original thought.
And then… it didn’t do any of those things. Not for the people using it well.
So now the goalpost moves to watermarks. “See? They’re MARKING it. That proves it’s bad.” No. It proves it exists.
I use AI almost every single day. I talk about it publicly. I teach people how to use it. I built a business around it. And not once has AI replaced my thinking, my voice, my stories, or my ability to look a client in the eye and stand behind the work.
AI made me faster. More precise. More prolific. AND more creative, if I’m being honest. The work is still mine, because I show up with something to say before I ever open a chat window. That’s the difference.
It's not about whether you use AI, but rather about whether you have anything worth saying without it. The watermark doesn’t change that equation; it just makes it harder to fake, and I'm fine with that.
Curious to hear your thoughts.