The latest buzzword is “AI slop.”
And honestly, I think it is a useful one.
But AI slop is not simply “content created with AI.”
That misses the point.
AI slop is what happens when someone treats the first or unfinished AI drafts as “good enough” and publishes it as the final version.
Decent but unfinished content is not the antidote to procrastination. Before AI, and now with AI, sloppy work is sloppy work.
We are quickly approaching a time where, in a professional environment, people can tell the difference between bare-minimum AI use and thoughtful AI use.
You have probably heard of the uncanny valley problem with AI images, where something looks almost right, but still feels off.
I believe something similar is starting to happen with AI writing.
That same intuitive sense that tells people an image was generated by AI also starts to work against the person who presents unedited AI output as their own thinking.
Not because they used AI.
Because they did not add enough meaningful human judgment.
And that judgment can show up in a lot of ways:
- Skills.
- Project instructions.
- Saved context.
- Memory.
- Advanced prompts.
- Better feedback to the LLM.
- Clearer examples.
- More specific direction.
Thoughtful AI use creates better content, deeper meaning, and a sharper perspective.
My rule is pretty simple:
👉 Treat all original AI output as a draft, and accept that it will rarely, if ever, be ready to publish after the first prompt.
Always a draft.
Here are five warning signs I look for:
1. It sounds right, but says nothing
If you can delete the sentence and the meaning does not change, cut it.
2. There is no point of view
If anyone could have written it, no one will remember it.
3. It feels like a remix
AI is very good at summarizing what already exists.
Your job is to add the experience, the example, or the opinion.
4. It is over-polished, but under-human
Perfect grammar does not equal trust.
Sometimes the post needs a shorter sentence.A rougher line.A little more of you.
5. The formatting gives it away
I recently had a client where employees were leaving the em dashes in their AI-assisted content.
Not once in a while. Everywhere.
And once you see it, you cannot unsee it.
Too many em dashes.Too many perfectly balanced bullets.Too many phrases like “unlock the power of” or “in today’s fast-paced world.”
Those are tells, not because AI used them, but because no one cleaned them up.
The answer is not to avoid AI.
Before you publish anything AI-assisted, ask yourself one question:
What did I add that AI could not have known on its own?
Did you add context from the business?
A real example?
A decision you made?
A lesson from experience?
If not, you probably do not have finished work yet.
You have a draft with better grammar.
And that is usually where AI slop begins.