Given the choice between an AI doctor that's 80% accurate and can explain its diagnoses and one that's 90% accurate but can't, I'd pick the latter.
-
I think you are assuming the 90% accuracy is aleatory. If it simply makes errors on all people named Pedro, that wouldn't be acceptable would it?
-
-
But to discover the "Pedro" problem, you would need some form of transparency / testability. I think we can achieve that.
-
The role of explaining how it is done is overplayed. This article from the New Yorker has some good points on this: https://www.newyorker.com/magazine/2017/04/03/ai-versus-md …
Keskustelun loppu
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.