AI paper evaluations could worsen replication problems, one post warns
A post describes an experiment with AI systems revisiting published research, finding opportunities and issues, then sharing those judgments publicly.
TLDR
One post describes AI systems rereading published research and publicly sharing judgments about opportunities and issues, calling the experiment a sign of a "tsunami coming for academia." A quote-post warns that using machines as evaluators will worsen what it calls the already "terrible" state of replication across disciplines.
Combined views
5.4K
3 Sources, first seen 23d ago