Is It Cheating? Using AI at Work, at School, and in Your Own Writing

Did I do this?

The report took an hour instead of a day, and you checked it. The numbers are right and the recommendation is yours.

And still, with the cursor resting on send, half a second goes by. In that half-second you wonder whether the report counts as your work.

Keep the pause, because under it sits one of the real questions of this decade. Nobody has settled it, and both answers have serious people behind them.

Yes, it's cheating

When someone asks for your work, they often want more than the finished thing.

A teacher who sets an essay wants to know whether you've learned to build an argument. A reader who buys your book is paying for your mind. Hand over what a machine produced, and you've answered a question nobody asked.

Fairness comes into it too, because if you used AI and your colleague didn't, you were compared on different terms. Then the comparison means nothing, since a grade, or your name on a piece of writing, exists to tell someone what you can do.

Then there's the slow cost. When AI thinks for you, you don't. The report got done, but your ability to write reports got a little weaker.

Do that for ten years and the question changes from "did I cheat?" to "can I still do this?"

None of that is paranoia. The people making this case are describing something real.

No, it isn't

Nobody calls a calculator cheating now, or a spellchecker, a search engine, or the colleague you ask "how would you phrase this?"

Each was called cheating when it arrived. Each time the line moved, and it always moved the same way, towards the tool doing the mechanical part so the person can do the judging.

So what's the work? It's knowing what to ask, and telling a good answer from a bad one.

Typing was always the smallest part. If you briefed the AI, checked its answer, changed it and put your name on it, you did the work. You just did it faster. Your manager wanted the problem solved, and the problem is solved.

Fairness cuts this way as well, because everyone has the tool in their pocket.

Refusing AI doesn't make you more honest. It makes you slower than the person who used it well. In most jobs, "I could have done it in an hour but I chose a day" just loses you a day, and nobody calls that integrity.

And the people who hold this view aren't lazy either.

Ask what the work was for

Both sides argue about the same thing, but neither names it.

A more useful question is what this piece of work was for, and whether using AI defeated that purpose.

Some work is for the result. Think of the report that helps someone make a decision, or the email that gets the refund. Here the AI is just a tool, and nobody asked you to suffer.

Some work is for the practice. Think of the essay that teaches you to argue, or piano scales. Handing that to the machine is like paying someone to go to the gym for you, so you get a receipt and no muscles.

Some work is for the person. Think of a letter to your dad, or the speech at your best friend's wedding. The only thing being asked for is you, so a better paragraph that isn't yours is a worse gift.

Most work is a mix. Your report is mostly for the result, but the moment you decided what to recommend was for the practice. That moment is the one to keep doing yourself.

Using AI isn't cheating. Pretending you didn't use it, when the person asking would want to know, is cheating.

Letting AI do the part that was meant to make you better is a quieter kind of cheating, and its only victim is you.

My line is to use AI for the result, do the practice yourself and keep the personal things personal. And be able to say out loud how you made the thing.

The best case against me is "deskilling", which means slowly losing a skill because a tool does it for you. I don't think that case is wrong.

Spend ten years letting AI draft every report, and the person who could write one from nothing may not be there any more. My answer is to keep doing some of it by hand on purpose, like driving with the satnav off on roads you know.

Whether that's enough, I'm not sure anyone knows yet.

At work, at school, on your own page

At work the job is mostly for the result, so AI is mostly fine on two conditions. First, you've checked the work well enough to defend every line in a meeting. Second, your employer hasn't said no.

Some employers have, usually because of worries about their data, and "I didn't think the rule applied to me" won't defend you. Would you squirm saying "I drafted this with AI" out loud? If you would, find out why before you send the work.

At school and university the work is mostly for the practice. The line sits tighter, and the school or university draws the line for you. Find out where your school draws the line, because the rules differ from place to place.

Honest use is what the study page describes. There, the AI quizzes you and marks you, but it never produces the thing you're meant to be able to produce.

If the assignment is "show you can write an essay", an AI essay is a different assignment.

In your own writing the work is for the person, so the call is yours. But be honest with yourself about what the reader is buying.

If the ideas are yours and the AI just tidied the sentences, that newsletter is yours. If you type a topic and post what comes out, it's the AI's newsletter with your photo on it, and readers can tell sooner than you'd think.

Think of the last thing you used AI for, and the half-second before you sent it.

What was that work for, and did the way you used AI serve it or get round it?