AI Slop, Suspicion, and Writing Back
The impetus for this post was my recent realization that I’ve developed an involuntary reflex for spotting AI-generated content. The tells are subtle now, but (sadly? tellingly?) this sort of content is seemingly everywhere now once you start looking.
One bit of hipster cred I get to claim is that I followed Simon Willison before he became the cool AI blogger.1 Perhaps one of the most “in the room” feels I’ve had reading his work is to see his neologism of “AI slop” catch on.2 Slop being defined as the equivalent of “spam”, but for AI-generated content.
For example, prompting Claude to write a blog post and publishing that verbatim under your name would be “slop”, even if the writing is not bad at face value.
GPT-3 era slop was pretty easy to detect. Early GPT-4 era slop was marginally more convincing, but usually gave itself away after a paragraph or two. Sometime in 2024, I think some critical line was crossed, at least for me, and there are certain classes of slop that take at least some critical thinking to realize is AI-generated. Which isn’t great!
