Tootfinder

gen-AI outputs tend to be locally plausible but globally incoherent, because there's no intention, no viewpoint that motivates the assembly of the parts.
with a little practice, people can get pretty good at sensing "the vibes are wrong", and from there pick out specific details that are suggestive of gen-AI. note that this process goes from global to local, but the explanations tend to be just the local, because it's hard to explain "the vibes are wrong"