By Susan Sly | August 2026
Quick answer: If you create content, run a brand, or use AI anywhere in your business, the new AI watermarking protocols will affect how your work is seen, ranked, and trusted. This article explains what AI watermarking actually is, what it means for your AI search visibility, and the deeper question almost nobody is asking: can a watermark even tell whose voice is whose anymore? It will take you five minutes, and it may change how you publish everything.
Before we get to the technology, I want to tell you where I was standing when this news found me, because the setting turned out to be the whole point.
A room full of people finding their voice
This past week I was in Los Angeles, volunteering at Mastery in Transformational Training, the leadership development program I graduated from. Chris Lee, the lead trainer and a coach to executives and celebrities alike, was masterful. I watched him help a room of emerging leaders break through the limiting beliefs that had quietly run their lives for decades.
With Chris Lee at my MITT Basic graduation, March 2025. This past week I returned to the same training as a volunteer.
Volunteers keep their phones in the volunteer room, so I was fully present. And in that presence, I found myself contemplating my own former limiting beliefs. Lack of confidence. Low self-worth. The quiet question I once carried: at 53, do I still have a voice?
A lot has changed since my own graduation. This year I have spoken on AI for Fortune 500 companies. I am launching my learn on demand program, The Get FOUND Method. We launched Amsara Health. We took hard feedback from our test users and rebuilt the Harmoni app from the ground up. I have a waiting list of consulting clients. And using my own techniques, I ranked number one in AI search for female AI inspirational keynote speakers.
I share that list not to impress you, but to make a point about what happens when a person finally claims her voice: everything downstream changes.
Then, on the break, I checked my phone. The AI news of the day stopped me in the hallway.
The news: your AI now signs its work
The major AI companies have begun embedding watermarks in AI-generated text. Anthropic now includes an invisible, machine-readable watermark in text generated by Claude, part of its compliance with the EU AI Act. Here is how it works in plain language: as the AI writes, it subtly favors certain word choices from a predetermined set. No single sentence reveals anything. But across hundreds of words, the pattern becomes statistically unmistakable to anyone holding the detection key. Google has a similar system called SynthID. Analysts covering the rollout describe it as a signature woven into the fabric of the text itself.
Standing in that hallway, I had two immediate questions. What does this mean for my clients? And what does it mean for the AI search optimization techniques I teach?
What watermarking means for your AI search strategy
The honest answer is that the evidence is mixed, and anyone who tells you otherwise is selling certainty they do not have.
One early study found that heavily watermarked AI content ranked lower in search and was cited about half as often by AI answer engines. Other analysts argue the impact is negligible, because search engines cannot independently detect the watermark without the AI company’s key, and because Google has said for years that it rewards quality, not origin.
Here is what both camps agree on, and it is the part that matters for you: the watermark weakens dramatically under heavy human editing. When you rewrite, restructure, and add what only you know, the statistical signature dissolves into something new. Something yours.
Which means the strategy I have taught all along does not change. It gets confirmed. AI is a drafting partner, not a ghostwriter. The content that wins in AI search is the content only you could have written: your stories, your data, your Tuesday. I wrote about this principle in my honest comparison of Claude Fable 5 and GPT-5.6 Sol: use the frontier models for work that carries your name, and then make certain your name deserves to be on it.
The paradox nobody is talking about
Now for the question that kept me thinking long after the break ended.
At the same time watermarking arrived, ChatGPT rolled out memory features that reference everything you have ever told it, learning your phrasing, your preferences, your patterns, so it can write more like you. One system is stamping text as machine-made. The other is training the machine to sound exactly like you.
So I will ask the question directly: can we have watermarking when AI is allegedly writing in your voice?
If an AI has studied thousands of your sentences and produces a paragraph you would have written yourself, the watermark says “machine.” Your reader hears “you.” Both labels are partly true, and neither is the whole truth. The technology can detect patterns. It cannot detect authorship in the way that actually matters: who lived it, who believes it, who stands behind it.
Whose voice is it, really?
That afternoon in Los Angeles, I watched people twice my confidence deficit and half my age wrestle with the same limiting belief I once carried: the belief that their voice did not matter. And I realized the watermarking debate is asking a version of the same question. Whose voice is this? Does it count? Is it real?
Here is what I know after this year. A watermark can identify which words a machine suggested. It cannot manufacture a story about standing in a hallway at 53, reading AI news between sessions of a leadership training you once sat in as a student, realizing every limiting belief you dropped made room for something you built. No model generates that. No protocol detects it. It is simply yours.
So use the tools. I use them every day, openly, and I teach my clients to do the same. But the work that carries your name must carry your voice, your edits, your experience, and your conviction. The watermark question will keep evolving. The voice question is already settled, and it was never about the technology.
You have one. Use it.
Susan Sly is an AI keynote speaker, consultant, and the founder of The Pause Technologies, makers of the Harmoni® app and parent company of Amsara Health. She is a graduate of executive programs at MIT Sloan and the MIT School of Engineering, and the creator of The Get FOUND Method. Book Susan to speak at susansly.com.
