I use AI to write and express ideas. Does that make the results SLOP?

There is a small ritual I do before I publish anything. I read back through the text and replace every em-dash, one by one. The long dash that almost nobody actually types on a keyboard has quietly become a watermark, a tell, the thing people point at and say "ah, a machine wrote this." So I swap them out for commas and full stops and the occasional pair of brackets, and I tell myself the writing now looks a little more like it came from a person.

I should be honest about why I do that. I use a large language model to help me write. I have ideas, lots of them, but they arrive scattered and out of order, more of a pile than a thread. I am dyslexic, and the gap between what I am thinking and what I can get onto a page has always been wide. Word choice and sentence construction are things I agonise over for far too long, and the agonising rarely improves the idea. It just delays it, or quietly kills it.

So here is the question I keep circling. When I work through an idea in conversation with an LLM, read the result, push back on it, edit it, and shape it until it says what I actually mean, is the result lesser? Is it slop? Or does it deserve to sit alongside the work of someone who writes fluently but whose ideas, frankly, are often weaker than mine?

The honest answer starts with noticing that we constantly confuse two different things. There is the quality of an idea, and there is the quality of its expression. We like to pretend these are the same, that clear writing must reflect clear thinking, but anyone who has read a beautifully written piece that says almost nothing knows better. Fluency is persuasive. It buys trust it has not always earned. A polished sentence can carry a thin thought a very long way, and a clumsy sentence can bury a good one before anyone gives it a chance. Writing has always been a gate, and plenty of people with something worth saying have been turned away at it.

It also helps to remember that writing was never the lone heroic act we like to romanticise. The image of the author at the desk, summoning perfect prose from nothing, was always partly a flattering story. Real writing has copy editors, spellcheck, grammar tools, a friend who reads the draft, a thesaurus open in another tab, dictation software, ghostwriters for the people who can afford them. We long ago decided that all of that was fine. Nobody accuses you of fraud for running a spellcheck. So the interesting question is not whether I used a tool. It is where, exactly, we have decided to draw the line, and whether that line is about quality at all or just about how new the tool is.

For me the line gets even harder to defend, because the friction I am removing is not a noble struggle that sharpens my thinking. It is a barrier sitting between my thinking and the page. People without dyslexia sometimes assume the effort of writing is where the good ideas get forged. For me the effort mostly happens before any of that, in my head, and the writing is the part where the idea gets lost on the way out. The thought is already there, fully formed and worth having. What the tool does is stop it dissolving somewhere between my head and my hands.

This is where the word "slop" actually earns its meaning, and it is worth being precise about it. Slop is not defined by the presence of a machine. It is defined by the absence of a person. It is text generated at volume with nobody home, nobody reading it back, nobody deciding whether it is true or whether it is any good, content produced because content is cheap and attention is profitable. The defining feature is that no one cared. If I have wrestled an idea into shape, argued with the draft, cut the parts that were merely fluent, and stood behind the result, then whatever that is, it is not slop. It might be wrong. It might be boring. But someone was home.

I want to be fair to the strongest objection, because I feel it too. There is a school of thought that says writing is not just how you express a thought, it is how you discover what you actually think. The struggle to find the right word is where vague feelings get forced into precise claims, and if you hand that struggle to a model, you risk handing over some of the thinking with it. I think that is true, and it is exactly why the editing matters more than the drafting. The conversation is not me asking for an answer. It is me recognising what I do and do not mean once I can see it written down, which is something I have always done far better than I can manage cold.

And I will not pretend the easy button is not seductive. The real danger of this tool is that it makes stopping feel like finishing. There is a competent, frictionless paragraph available at any moment, and the pull to take it and move on is constant. Let it win, and the model hands me something smooth and slightly hollow, and I decide it is good enough, and good enough is how you publish things you never actually thought. That would be slop, and it would have my name on it. So the work has not gone away. It has moved. It used to be getting words onto the page at all. Now it is refusing the easy ones.

So I am left with the two accusations that actually sting, the ones hiding underneath the word "slop." Am I cheating? Am I stealing?

Cheating implies rules, an exam I was supposed to sit unaided. There is no exam. There is only a romantic idea of how writing is meant to happen, and I never signed up to it. I am not pretending these words cost me what they cost a fluent writer. I am telling you plainly how they got here, and you cannot cheat at a game you are playing in the open.

Stealing is the more interesting charge, because it asks whose work this really is. The part that matters was never the model's to give. The ideas are mine. The judgement about what is true and what is rubbish is mine. So is the choice to keep this sentence and cut that one. What I took from the model is phrasing, the same thing any writer takes from an editor or a thesaurus, and nobody calls that theft. And if I am dismissed for how the words reached the page while a fluent writer with a thinner point is waved through, that is not really a standard. It is just a preference about process.

I am not ashamed of any of it. The thinking is mine. I just finally found a way to get it onto the page intact. Whether the thinking was any good is a fair question to ask. It is also the only one that was ever worth asking.