What is AI slop?
AI slop is generated content published with little or no human judgment applied, recognisable because it is fluent, confident, and says nothing that could only have come from the person whose name is on it.
Also called: AI generated content spam, generic AI content, AI writing tells.
The tells are consistent. Openings that set a scene without making a claim. Lists of three where two would do. Sentences built on "it is not just X, it is Y". Hedged conclusions that commit to nothing. None of these is wrong in isolation; the pattern is what gives it away.
The underlying problem is not the model, it is the absence of input. Asked to write about a subject with no specifics to work from, any writer produces the average of everything already written about it. Readers recognise that average immediately, because they have read it many times.
Which points at the fix. Content stops being slop when it carries something only the author has: a number from their own work, a decision that went badly, an opinion someone could disagree with. Generation is not the problem, publishing without adding anything is.
The reputational cost is asymmetric and lasting. One obviously generated post teaches an audience to skim everything you publish afterwards, and that is expensive to undo.
Common questions.
How do I stop AI-written posts sounding generic?
Give the draft something only you have: a real number, a client reaction, a decision you got wrong. A tool with nothing specific to work from will produce the average of everything written on the subject, which is exactly what readers recognise.
Can readers really tell when a post is AI-written?
They can tell when nobody added anything. A draft generated from your own writing and then edited reads as yours; one published untouched reads as the average of the internet.
In the product
See also