“Now we’ll see who writes with AI.”
Good. Let it show.
Anthropic has announced that it marks texts produced by Claude with an invisible signature: it stays inside the text, survives copy and paste, and is there to comply with the European transparency obligations that came into force on 2 August. I wrote about it at the end of July, when they were still a line of regulation. Now they’re a label.
Within days it became the topic of the moment, and the prevailing tone is alarm.
The concern is understandable. Those who write for a living fear being confused with the machine; those who write for work fear seeing their output devalued. It’s a legitimate fear.
But a label says what it says — not what we’re afraid it says. The watermark doesn’t certify who thought. It certifies where the text passed through.
Those are two different things, and we already know a version of this.
On the back of every iPhone it says: “Designed by Apple in California. Assembled in China.” Two lines, two crafts, no shame. Apple never hid the second one: it keeps it right there precisely because it’s confident about the first.
I’d sign my text the same way. Proudly designed and created by me, written with AI.
That’s not a pre-emptive excuse. It’s the opposite. I’m telling you that what I think holds up well enough for me to tell you how I produce it.
Because the part that matters stays where it’s always been. AI doesn’t know which problem is worth tackling. It doesn’t know what I saw at a client meeting on Tuesday. It doesn’t know when an idea is spent and needs cutting. And it has no idea what keeps me awake at night when I have a problem to solve.
All of that goes into the first line. The watermark marks the second.
Fill the first. Write the second openly, and write it large — because nobody else can write the first one for you.