“Explainability” Is a Poor Band-Aid for Biased AI in Medicine
Ғылым және технология
New study shows docs not great at sniffing out bad models.
Click here to subscribe: kzread.info...
Get more medical news analysis at www.methodsman.com or www.medscape.com
#AI #medicine #science
Пікірлер: 6
this needs more views. There is too much hype and optimism, and not enough skepticism around AI. It is not even close to being able to be trusted in medicine. The current (and near future) state of AI and AI hype is like that of Tesla Fool Self Driving.... You can dream about the perfect self driving (if it's even achievable), but the current state of FSD is not even close to being ready for primetime.
Now I'm waiting for a study on training doctors in recognizing bad training in AI. Talking from a different industry where AI generated content is a big percentage of my work, this seems very familiar, and in my case it's up to me to figure out if what AI generated makes sense or not. I find that the less I know about certain topic the more I'll trust AI to do the work for me. I hope the models for medical purposes will be vetted by pros.
Congratulations on your ASCI! Well deserved.
Impact of Bad AI is more than the impact of Good AI ! But not sure if it is an apple to apple comparision. With good testing, and safety protocols, I'd believe we can limit the access and deployment of Bad AI !
Interesting. Thank you for sharing
Loved the plaid shirt, but please skip those nonsensical turns. As always, your info was hugely relevant and clearly presented. Thank you!