Post #2386854
2026-04-14 01:23 UTC
@eljojo@ruby.social Not so long ago, the proportion of false positives and recalls were significantly higher with automatic detection than by an experienced clinician, but this study seems to indicate that we've reached a point where it's roughly as good. People have been doing research and looking for this level of performance for a couple decades by now.
Replies (1)
-
@eljojo@ruby.social 2026-04-14 15:42
@ooze@cathode.church that’s some really good context. I think I read somewhere that, in some contexts, we’re trending towards AI alone being better at diagnoses than AI + Human, due to bias reducing accuracy.