Similarly, "AI Sci Slop" (or "Sci Slop" for short) is a term I've coined not for simply underdeveloped AI and related tech for science, but specifically as a taxonomic category for unvalidated and overhyped (overt or by subvert association).
I've been using the term and its has made some people mad, with
@BioProtocol even accusing me of "defamation" (of AI?!?).
I'll be rolling the term out in a first formal forum at my panel on DeSci at the College of American Pathologists conference next week, and peppering it around at the Department of War's "Medical Autonomous Care State of the Science" meeting I've been invited to for my DeSci/AI perspective.
I'm also conducting a self funded (because DeFiDeSci gatekeepers wont fund it, heh) independent assessment of some tools out there like
@Aubrai_ to get a better alignment with the current offerings of tech for science with standardized scales like NASA's tech readiness level - akin to what Fleuren et al did for medical AI.
In science, you can just do things - though you also want to define your terms, describe your methods, present findings, gather and incorporate feedback, etc.
‘AI slop’ isn’t slop because it’s mechanically produced, it’s slop because it’s statistically produced. It’s similar in nature to design by committee, the focus group, data-driven web design, etc.