Notices by jumblies (jumblies@mastodon.social)

@mikez_petm disease on medical images is bullshit. It can identify a very small subset of pathology and with poor accuracy.

@mikez_petm as with all AI claims, the burden of proof is not on those refuting the claim, but on those making the claim.

There are plenty of research examples of AI working in radiology including some that proof of concept is on synthetic data. There are however, few examples ready for a production healthcare environment and those that are are vastly inferior to humans.

I'm being factual, not a naysayer.