The Practice Room / AI / Episode 8

AI

Is AI slop the machine’s fault?

By MstimajPublished 17 November 2026Level AI FluencyTool ChatGPT

Sign in to save your place and mark this done. Free, no card.

What she says in the film

The short answer

No. Not one of the four things that make something slop is the machine’s fault, and that includes the volume.

Slop is a word that refers to food. It is what you tip in a trough, and four rules qualify something as slop: it is shaped like a meal, nobody cooked it, it comes by the bucket, and the cook does not eat it.

All four fit what gets posted. All four describe a person.

The four rules, under the AI umbrella

  • Shaped like a meal becomes shaped like content. The right length for the platform, the right tone, a hook at the top. It looks like a video. It looks like writing.
  • Nobody cooked it becomes nobody made it. Nobody wrote that. Nobody filmed that. The machine did, start to finish, and it went out exactly as it came back.
  • It comes by the bucket becomes mass quantities, on command. This is the reason the word stuck to AI at all.
  • The cook does not eat it becomes nobody read it back before it left.

Why the volume is yours too

The bucket is the one people hand to the machine, and it does not belong there. A tool has never decided how often you publish.

You set the cadence. You decide whether today is one post or four, one film or three. Every piece of volume in your feed was authorised by a person who could have said no. The machine executes the quantity. It does not choose it.

Do it yourself

  • Open the last thing you published. Not somebody else’s feed. Yours.
  • Run the four tests on it, honestly. Shaped like content? Did anyone actually make it? Did it arrive as part of a batch? Did anyone read it back?
  • For every yes, write down which of your decisions caused it.
  • Look at your posting schedule and ask who set that number. If the honest answer is a growth tip you read once, that is a decision you can take back.

Check yourself

Is it slop just because AI touched it?

No. The tests are about whether anyone turned up, not about which tools were open. The same tool in a different pair of hands produces work nobody would call slop, and you have seen that work without labelling it.

But the machine really can produce more than a person can.

It can, and that is the one honest part of the complaint. What it cannot do is decide to. Capacity is not a decision. Somebody still types the number into the calendar.

What is the quickest way to stop making it?

Bring a draft, learn the subject, bring what you already know, curate, review every line, then build the thing you could not have built alone. Six moves, and not one of them is a prompt.

When it goes wrong

The usual mistake is treating this as a verdict on other people. Run on somebody else’s feed the four tests are just ammunition. Run on your own last post they are useful, because every one of them points at a decision you can make differently tomorrow.

The second mistake is hearing all four are you as all four are your fault. They are not a moral judgement. They are a list of the places where you still have control, which is the opposite of the argument that says the machine did this to us.

Words used here

Go deeper

The long form runs the whole argument: where the word began in a kitchen, how the four tests map onto what gets published, and the six-move method for not producing any of it.

That is the last episode of AI so far. New ones land as they are filmed.

On your machineWant this set up on your own computer, with your own files?

Book a session and we do it together on a call.

Book an hour

Ask what did not make sense

Comments go through our own form and are read before they appear. Links are not allowed; describe it in words.

Comments are read before they appear.

The newsletter

Get the next walkthrough.

One email when something goes up. The argument, the film and the term, nothing else. One click to leave.

Have a question now? Ask Mstimaj, she knows the archive.