The Big Fuss over Claude Watermarking Missed an Important Fact

It's interesting to see all the posts about Anthropic/Claude adding watermarks to text and image content, because all the fuss is missing an important fact:

Google's been doing this (almost) all along.

They call it SynthID:

"SynthID embeds digital watermarks directly into AI-generated images, audio, text or video. The watermarks are embedded across Google’s generative AI consumer products, and are imperceptible to humans – but can be detected by SynthID's technology."

(OpenAI/ChatGPT adopted SynthID, too, to watermark both AI-created images and audio.)

It begs the question: If AI can identify other AI-written content, why would they use your AI content in an answer when they can just create their own AI-generated answer? Wouldn't they (as Google has said countless times) prefer content that has actual human experience, expertise, etc.? (Hint: Yes, they would.)

I did a webinar two years with one of the major real estate website platforms and most everyone looked at me like I was an alien when I said "you have to un-AI your AI content." ¯\_(ツ)_/¯

This is why I've been banging the drum since day 1 for "AI-assisted" content -- the AI is your assistant, not the author. If you think publishing 10x more AI articles, videos, podcasts, etc., is the path to longterm success, I have a feeling that's not going to end well for you.

If you liked this article, you'll love my newsletter. Marketing Unlocked arrives in your inbox every other Sunday at 6:00 a.m. ET. You can unsubscribe anytime, but 500+ of your industry peers are reading it and raving about it. Click here to learn why and subscribe.