Is This Cheating? An Honest Answer From Someone Who Uses It

Someone asked me this at a dinner in a tone that made it clear they had already decided, and I gave a bad answer, because the honest one takes longer than a dinner table allows. So here it is properly.

I do not think people who object to this are stupid or frightened. Several of their objections are correct. I also do not think the question has one answer, because “cheating” only means something relative to a game you agreed to play.

What the tool actually removes

Execution. That is the whole of it. The ability to render what is in your head onto a surface with your hands, which is a skill that takes years and which I do not have.

That is not nothing. It is a large and beautiful thing to be able to do, and people who spent a decade acquiring it are entitled to feel strange about a text box producing a passable version in forty seconds. Anyone who waves that away is not arguing in good faith.

What it does not remove

Everything upstream and downstream of execution, which turns out to be most of the work.

Deciding what is worth making at all. Noticing that the thing you made is boring, which is a much harder skill than it sounds and which the tool actively works against, because a generated image is always technically competent and therefore always slightly flatters you. Knowing what to change. Knowing when to stop.

I write about the craft of prompting as though it is a technique, and mechanically it is, but the reason two people using identical instructions get different results has nothing to do with the instructions. It is that one of them can tell when an image is dull and the other cannot. No prompt guide teaches that, including ours.

The objections that are correct

Training data. These models learned from an enormous amount of work made by people who were not asked and were not paid. That is a real argument about consent, and the fact that it is legally unsettled does not make it ethically settled. I use the tools and I do not think that argument has been answered. Both of those things are true at once.

Displacement. Illustrators are losing commissions. Not in a speculative future, now, and mostly at the low end where people were making a living from work nobody considered precious. Anybody claiming this is only creating opportunity is not looking.

The flood. The cost of making a picture fell to near zero and the amount of pictures rose accordingly, which makes every individual picture matter less. That is a genuine loss and it is not fixable by anyone being more careful.

The objections I do not accept

“It has no soul” is an aesthetic claim dressed as an ethical one. Some generated images are lifeless and so are most photographs and most drawings; the ratio is worse here because the barrier is lower, not because the medium forbids anything.

“You did not make it, the machine did” proves too much. The camera made the photograph too, in exactly that sense, and we settled that argument a century ago by deciding that choosing the shot is the authorship.

“It is easy” is true of the first image and false of the fiftieth. Getting something specific rather than something generic is genuinely difficult, as anyone who has tried to get the same face twice will tell you. It is a different difficulty from drawing, and much smaller, but it is not zero.

So: is it cheating?

It is cheating if the thing being evaluated was your ability to draw. If someone asked you to make a drawing, generating an image is not an answer to that question, and passing it off as one is straightforwardly dishonest.

It is not cheating if what was wanted was a picture. Nobody accuses you of cheating for hiring a photographer, and the moral content of the act does not change because the photographer is software.

The interesting cases are in between, and the test I use is whether anyone would feel deceived on finding out. A wallpaper, a mood board, a thumbnail nobody thinks about: no disclosure needed, nobody was assuming a person sat and drew that. A portfolio, an illustration in a news piece, anything where the making is part of what is being valued: say so. Not because a rule requires it, but because the discomfort you feel about mentioning it is itself the answer to whether it needed mentioning.

What I actually think

I think this is a tool that removed one barrier and left every other barrier standing, which people consistently misjudge in both directions. The people who think it makes them artists are wrong. The people who think it makes art impossible are also wrong.

And I think if you feel slightly uneasy about it, that instinct is not naive. It is responding to something real about the training data and about the people whose work got harder this year. Using the tool anyway is a defensible position. Using it while pretending those costs do not exist is not.

If you want the practical side rather than the argument, the first-week account and the honest cost breakdown are both a lot less philosophical, and the studio asks you no questions at all.