The Trouble With Calling Everything AI Slop

The awkward truth about AI writing: people may like it more than they think

‘AI slop’ has become one of those handy little phrases people love throwing around.

And to be fair, some of it deserves it. We have all seen that strange, shiny, lifeless writing that sounds perfectly confident while saying almost nothing at all. It is polished in the same way a hotel corridor is polished. Clean enough, but you would not choose to spend time there.

So yes, AI slop is real.

But here is the awkward question no one seems quite as keen to ask.

What if people actually like AI writing more than they think they do?

That is what made the recent New York Times quiz so interesting. Readers were shown different writing samples and asked which they preferred. It was a blind test, so no one knew which passages were written by humans and which by AI. Just words on a page, doing their best. Or in some cases, doing their absolute worst.

The result was the bit that made people sit up a bit straighter. After 86,000 people had taken the quiz, Kevin Roose shared that 54% of the votes had gone to the AI-written passages. So most people preferred the content written by AI, than that written by human beings.

That does not mean we should all light a candle for the death of human writing and quietly back away from our keyboards. It does not mean AI has become the next great novelist. It does mean something a little more awkward and a lot more interesting.

People may be more biased than they realise.

I speak to lots of people that say they can always spot AI writing. Additionally lots of people also say they would never prefer articles written by AI. But once the label disappears, some of that certainty seems to fall apart rather quickly. In this case, enough readers chose the AI version for it to come out ahead.

Which suggests the problem is not always the writing itself. Sometimes it is the reaction people have to the idea of it.

That is where the conversation usually goes a bit wrong.

Too many people talk about writing as though there are only two categories. Human writing, which is deep and wonderful and full of soul. Or AI writing, which is bland nonsense that should be escorted off the premises.

But that is not really how it works, is it?

Plenty of human writing is dull. Some of it is just awful. There are emails, reports, and website pages all over the country right now that could send a healthy adult into a small nap. Humans are not automatically brilliant just because they have a pulse and access to a keyboard.

And AI is not automatically terrible just because it helped write something.

The real difference is often much simpler than people want to admit.

AI writes badly when people use it badly.

That is it.

If you give it a vague prompt, you get vague writing. If you give it a boring instruction, you get boring output. If you ask it to write something ‘professional and engaging’, you are essentially begging for a paragraph full of the sort of phrases nobody has ever said out loud with a straight face.

That is where the slop comes from.

Not from AI itself, but from lazy prompting and even lazier editing.

Used properly, though, it can be a different story. Give it a clear brief. Tell it who it is for. Explain the tone. Ask for specifics. Push it to drop the waffle. Make it sound more natural. Challenge the first draft. Add your own thinking, your own judgement, and your own point of view. Do that, and the output gets much better very quickly.

That is why prompting matters so much.

It is also a big part of what we teach. Prompting is not some daft trick for impressing people at networking events. It is the practical skill of giving better instructions so you get a better result. Do it badly and you get beige nonsense. Do it well and you can get writing that is clear, useful, and genuinely enjoyable to read.

Which brings us back to the New York Times quiz.

To me, the really important lesson is not that ‘AI beat humans’. That is a bit too dramatic and not especially helpful. The more useful lesson is that AI can produce writing people genuinely prefer when it is handled properly. And many people still underestimate that because they are judging the source before they judge the words.

That is the bit worth sitting with.

Because once you strip away the label, people do not always react the way they expect to.

So yes, let’s keep calling out AI slop when we see it. Some of it is dreadful and deserves to be gently but firmly launched into the sea. But let’s not pretend every bit of AI-assisted writing is doomed to be soulless drivel. Sometimes it is bad. Sometimes it is fine. And sometimes, apparently, people prefer it.

Which is awkward if your whole personality is built around saying you can always tell.

If you’d like to learn how to prompt AI properly and avoid creating content that sounds like it was written in a beige meeting room, you can find out more about our training here.