Writing off everything a machine makes as slop is as lazy as the slop itself.
"AI slop" has become the insult of the moment. A clumsy email. A soulless article. A LinkedIn post that says nothing in 200 confident words. Slop. Blame the machine.
It is a satisfying word. It is also a lazy one. Writing off everything AI touches as slop is about as considered as the slop it complains about.
Here is the thing we keep missing. Quality has always been a scale. That was true long before AI arrived. Plenty of human work is slop. The committee report nobody reads. The marketing email built entirely from filler. The blog post written to hit a word count rather than say something. We did not need a machine to produce hollow content. We have been doing it by hand for years.
AI sits on the same scale. The output runs from genuinely poor to genuinely excellent. What decides where it lands is not the tool. It is the person holding it.
Same model, two users, two completely different results. One types a vague request and pastes back whatever comes out. The other brings context, a clear brief and a point of view, then reads the output properly before it goes anywhere near the public. The first produces slop. The second produces something worth reading. The tool did not change. The operator did.
Take a simple example. Two social media posts.
The first is generic AI output and you can spot it across the room. It opens with "in today's fast-paced world". It is relentlessly upbeat. It makes three points that are really the same point. It has no specifics, no opinion, no risk. It could have been written for any business in any industry on any day. It says nothing, pleasantly.
The second is written by someone who knows their subject. It has a clear view. It uses a real example. It leaves things out. It sounds like a person, because a person shaped it, even if a machine helped draft it. You finish it with a thought in your head that was not there before.
Both might have used AI. The difference is not the tool. The difference is that one person treated the output as a finished product, and the other treated it as a first draft to be directed, cut and sharpened.
This is where the popular framing goes wrong. People talk about good AI work "passing as" human, as though the goal is disguise. That misses the point. The good post is not pretending to be anything. It is simply good. Provenance is not the tell. Quality is. Nobody asks whether a designer used a computer. They ask whether the design is any good.
There is a comfort in the blanket dismissal, of course. If all AI output is slop, you never have to learn to use it well. You can fold your arms and feel superior while the world moves on. That is a pleasant position. It is not a durable one.
So what happens next is not that AI makes everything slop. It is that the scale gets wider at both ends.
The floor drops. Anyone can now generate fluent, confident, empty content in seconds, and plenty of people will. The tide of slop is real and it is rising.
But the ceiling rises too. The people who bring judgement, taste and subject knowledge now have a tool that lets them do more of their best thinking and less of the grind. Their work gets better, not worse.
The gap between those two groups is going to widen. And it will not map onto human versus AI. It will map onto considered versus careless. The careless will hide behind the machine and blame it for the results. The considered will use it and quietly pull ahead.
Which is why "AI slop" tells you more about the maker than the machine. Slop is not what happens when a person uses AI. Slop is what happens when a person stops thinking and lets the tool think for them.
The skill was always the scarce thing. It still is. The tools have just made that easier to see.
Mat Keys is the founder of newlens, writing here in a personal capacity.