It writes faster than you think, fills a blank page in seconds and makes you look sharp. So is the work still yours?
You fire up ChatGPT, Claude or whatever tool you found this week, and you tell yourself you're being efficient. Modern, even. But underneath there's often a quiet discomfort, a sense you might have crossed a line you can't quite name.
You're not imagining it. And I don't think a committee or a regulator is going to draw that line for you. You draw it yourself, and it moves depending on your intent, your honesty and what you're willing to own.
Here's how it tends to go. You sit down to write. You know what you want to say, but the words won't come, so you ask AI to draft it. It does, beautifully. You tweak a sentence or two and hit send.
And then the guilt creeps in. Did I actually write that?
That guilt is pointing at something real. You've handed over the thinking, the crafting and the wrestling with ideas until they come clear. Now you're not sure where you end and the machine begins. Whether you used AI matters less than what you used it for. Does the final piece still carry your intent, your judgement and your voice?
Most people treat this as a yes or no. You use AI or you don't, you disclose it or you hide it, it's cheating or it isn't. That's too simple to help anyone. I think there are four boundaries worth checking, and each comes with one question.
Ownership
Did you come to the table knowing what you wanted to say? Or did you hand over the thinking and hope the machine would work it out?
Using AI to put an idea you already have into words is a very different thing from asking it to come up with the idea. Picture a chef. Getting help writing up the recipe for a dish you've perfected is one thing. Asking the machine to invent the dish, write the recipe and tell you how to cook it is another. In that second kitchen you're following instructions.
Owning it also means you'll stand behind it. You've reviewed it, shaped it and made sure it sounds like you. The question: would I be proud to say I made this? If your honest answer starts "well, technically…", you probably don't own it.
Truthfulness
You don't owe anyone a breakdown of your creative process, and you don't need "partially generated by AI" stamped on every email. The test is simpler. Would someone feel misled if they knew the full story?
Say you're a consultant. A client hires you for your expertise and your ability to solve problems they can't. Using AI to structure your thinking, research faster or draft reports is fine, because you're still bringing that expertise. Feeding it the client's problem and handing back its answer with none of your own judgement is something else. They paid for you, and you've given them what a machine would have given them for free.
The question: would someone feel misled if they saw the process?
Impact on others
These models were trained on data that came from somewhere, and from someone. Writers whose work was scraped without consent. Artists whose styles get mimicked without credit. People whose faces turn up in synthetic media they never approved.
Most of us had no say in how the models were trained. That doesn't let us off the hook. You still choose whether to copy a specific person's style without their permission. You still choose whether to make a deepfake that could damage someone's reputation. And if you're automating work that used to employ people, are you thinking about the human cost?
The question: would I treat a person the way I'm treating this data?
Growth or avoidance
This is the one most people ignore, and I think it matters most in the long run. Are you using AI to go faster, or to avoid learning?
Speeding up tasks you already know how to do is smart. It frees up time for bigger thinking. Skipping the learning altogether is risky, because one day you'll hit a problem the AI can't solve and you won't have the skills to solve it yourself.
Take writing. If you can still structure an argument, craft a sentence and edit for clarity, using AI on a first draft is fine. If you never learned to write in the first place, you've got nothing to fall back on the day the tool lets you down. The same goes for coding, design, strategy and decision making.
The question: am I levelling up, or hiding? If you're hiding, it might be time to go and learn the thing you've been avoiding.
How it feels
There's a simpler test underneath all four. Does the work still feel like yours? Forget who typed the words. When you look at the finished piece, do you recognise yourself in it? It's not scientific. But you usually know when you've put yourself into something and when you've phoned it in.
Where does it show up? Brainstorming, outlining or polishing a draft is help. Publishing an article AI wrote, without adding a thought of your own, is where it tips into replacement. The same split runs through everything else. A melody you rework until it's yours, or a generated track with your name slapped on it. Research done faster, or AI's summary passed off as lived experience. Replies you personalise, or customers who never meet a human at all.
So what does using it well look like? You set the direction, with a clear angle and a point of view. You question what comes back. You check the facts and the sources. If something goes wrong, you take responsibility instead of blaming the tool. And you don't use automation to dodge difficult conversations or hard decisions.
If the tech vanished tomorrow, would your voice still stand?
I'd really like to know where you've landed on this. You've felt the tension, and maybe wondered if you've gone too far. What boundary have you set for yourself that you actually follow when no one's watching?
