Do I have to tell people I used AI to write this?

A reality check we all need to read.

You typed the apology before you typed the point. The email was good, the analysis was sharp, and yet your thumb hovered over a sentence that started with "I just used ChatGPT to pull this together, so apologies if".

You didn’t write that line for the reader. You wrote it for your own nerves, to soften the blow of being competent in a way that suddenly feels like it needs a permission slip.

The short answer is no, you do not have to disclose that you used AI to write something, and in most cases the disclaimer hurts you more than the AI ever could. This piece will show you why that little preamble is quietly costing you and hand you the one move that gets you out from under it, backed by the research that proves the bind is real rather than something in your head.

Why do I feel like I have to apologise for using AI?

You feel it because you have been quietly taught that effort is the proof, and AI removes the visible evidence of effort (that’s a crap belief).

For years, the sweat on your brow was how everyone knew the work was yours, so handing a draft to a machine and presenting the polished result feels, somewhere in your chest, like claiming credit you did not earn. That feeling is real, and it has a name, because the writer Rebecca Mbaya gave it one in her essay "The Apology Tax of AI Competence." The apology tax is the extra explaining, hedging, and disclaiming that competent people, women especially, add before the substance of their work, the small toll you pay for being good at something the room has not decided how to feel about yet.

Here is the part that should sting a little and then set you free…

The research finds that women arrive with longer disclaimers and more careful framing before they get to the actual point, and you have probably been doing exactly this in every email without ever having a word for it. That disclaimer is the tell. It announces your discomfort to the reader, and they hear the discomfort before they hear the work.

Is it true that people trust AI-assisted work less?

Yes, and the unfairness is sharper than you think, because the harshest judges are often the people who do not use AI themselves.

A study out of Hong Kong Polytechnic University and Peking University examined how identical AI-assisted work was rated and found that men who did not use AI judged women 26% more severely than they judged men for the same output. Identical work, scored differently depending on who produced it, with the gap falling hardest on the women in the room.

It gets more concrete. In that research, female professionals carried more than double the competence penalty for AI-assisted work, 13% against 6% for the men.

So if you have felt a low hum of suspicion that you were being scored differently for the same move your colleague made without a second thought, you were reading the room correctly. The unfairness is measurable, with a number attached to it, and it was never simply in your head.

What that data does not mean is that you should hide your use of AI or feel ashamed of it. It means disclosure is a strategic call you get to make on your own terms, as the leader you already are, rather than as someone confessing to a shortcut.

What is the no-apology send?

The no-apology send is exactly what it sounds like! You use AI to help with the work, then you present the finished thing with full ownership and cut the preamble that explains how it came together. You lead with your judgement and the result, and the method stays where it has always stayed, in your business and nobody else's.

Think about how you have worked for the last two decades.

THIS IS IMPORTANT…

When a sharp junior drafted a first version of a report, and you reshaped it and put your name on the final, you never opened the email with "apologies, Sarah did the first draft." You delegated it, applied your own judgement, and owned the result, because that is what leadership looks like.

AI sits in the same category as that junior. The delegation was never the betrayal, and the only dishonest part would have been pretending you did the whole thing alone.

So the next time your thumb hovers over "I just used ChatGPT to help with this," STOP! Read the sentence that comes after it. If the work is good, send the work. That apology never added a thing to the accuracy. All it ever advertised was your own discomfort.

But what if my workplace actually requires disclosure?

Then you disclose, and you do it as a leader setting a norm rather than someone confessing to a shortcut.

There is a world of difference between "sorry, I just used AI for this" and "I used Copilot to draft the analysis, and I changed the recommendation because I know the client's politics." The first hands away your authority. The second demonstrates exactly the judgement the machine cannot supply, and it quietly teaches everyone reading how disclosure is supposed to sound.

When you have to name the method, name the judgement in the same breath. Lead with the call you made and the context you brought that no model can reach. Framed that way, disclosure becomes a standard you are setting rather than a weakness you are admitting, and the whole difference is in who you sound like when you say it.

One thing to hold onto, though.

Always know your own organisation's policy, because some regulated industries genuinely require an AI-use note on certain documents. Following that rule is just doing the job well.

Apologising for following it is the apology tax creeping back in.

How do I actually break the apology habit?

Well, this is a habit I’ve been trying to break for years!

Start with your next important email, before it leaves your outbox. Write the whole thing, then go hunting for the apology, because it is almost always hiding in the first or last sentence dressed up as politeness.

Look for "I just," "apologies if," "this is rough but," and "I only used AI to," and delete every one of them.

Read what is left out loud.

If the work stands on its own, and it almost always does, you have your answer about whether the apology was carrying any real information or just carrying your nerves. Do this five times across a week, on five real sends, and the habit starts to loosen, because you will watch the world keep treating your competent work as competent work, exactly as it was always going to.

It’s a confronting test.

The shift here has nothing to do with confidence as a feeling. It is about removing one specific sentence from your writing until your natural authority has room to land. You spent twenty years earning that authority.

A two-year-old tool does not get to talk you into apologising for it.