Ancestors

Toot

Written by Anand R on 2025-01-12 at 17:24

This is incredible. https://kffhealthnews.org/news/article/artificial-intelligence-algorithms-software-health-care/view/republish/ Unsurprisingly it turns out what @pluralistic calls reverse centaurs are a terrible idea in medicine. After many paragraphs of calling out what a bad idea they are, the last 2 paragraphs say that the solution to failing "AI" can only be more "AI".

=> More informations about this toot | More toots from anand@mastodon.sdf.org

Descendants

Written by NixFREAK - on 2025-01-12 at 18:10

@anand @pluralistic the only “AI” healthcare should be research and development. AI general healthcare is a terrible idea “looking at you Mayo”

=> More informations about this toot | More toots from nixfreak@hackers.town

Written by Kate Nyhan on 2025-01-12 at 19:06

@anand @pluralistic

Thanks for sharing, very interesting article.

I don't think the last two paras are very pro the "more AI will fix bad AI" idea.... Certainly not the quotation:

“It’s great to have a vision where we’re melting icebergs in order to have a model monitoring their model,” Shah said. “But is that really what I wanted? How many more people are we going to need?”

=> More informations about this toot | More toots from kdnyhan@social.esmarconf.org

Written by teresa_athome on 2025-01-12 at 19:32

@anand

AI costs 20% more with a 35% error rate on life or death decisions and they want to add more AI 🤦‍♀️

@pluralistic

=> More informations about this toot | More toots from teresa_athome@mastodon.online

Written by Steve Popovich on 2025-01-12 at 20:06

@teresa_athome @anand @pluralistic because the more AI, the larger and more opaque the accountability sink becomes.

=> More informations about this toot | More toots from sspopovich@universeodon.com