New Slop furious
13 Poisoned Words on Reddit Can Make AI Search Lie
Cornell researchers showed a 13-word poisoned comment reliably steers ChatGPT and Google AI search toward scams.
Topic
Machine-made garbage flooding every feed you used to like.
Cornell researchers showed a 13-word poisoned comment reliably steers ChatGPT and Google AI search toward scams.
AI slop earned its own permanent beat, right next to crime and weather.
arXiv will bar researchers a full year for submitting machine sludge.
The recommendation engine fed kids roughly 40% machine-made sludge.