正在加载视频...
视频加载失败
"We should stop training radiologists now. It’s just completely obvious that within five years, deep learning is going to do better than radiologists." It's been 9 years since Nobel laureate and AI guru Geoffrey Hinton made this statement. Over those 9 years, my workload has only increased. Yesterday I... show more
11 条评论

Even if AI is very comparable to a well trained radiologist, we need expertise in PEOPLE PEOPLE can teach others and eventually improve AI Human + AI >> AI alone >> human alone, even when AI Is better than human

The college days are coming to an end soon. 🚀 Fun Fact: Our graduates are making serious bank! 💸 💡 HVAC: $32/hr = $66,560/year 🔌 Electrician: $29/hr = $60,320/year 🚽 Plumber: $27/hr = $56,160/year All from our 100% online courses that take just 100-120 hours to complete! 🤯 Most students finish in under 3 months. 🏎️ The ROI on these trades is 🔥 compared to college. Prove me wrong! 😜

Never forget that 50 yrs ago we were told to expect flying cars by now

Can AI help leverage your expertise vs. replacing it?

Get your laptop to embolize you liver mass or barbotage your shoulder and then get back to me Mr Coder. There are numerous studies proving coding and program bias in AI's. AI may change the way we practice but we we will not be redundant any time soon. "Clinical correlation required" will still be in the AI reports...

They are not getting it, the human experience is not this world, AI to assist cool but ai replacement? no thx. Radiologists are impressive. Urgent CTs save lives, imagine a hot stroke comes in and the AI crashes or fails? No one will be able to determine whether its an embolic stroke vs a massive bleed even if u can still do the imaging? No tpa and no chance at resolution of stroke symptoms. U r right, not safe. No one does RAD residency, eventually not one human on earth who cannot diagnose Xrays, CTs and MRIs manually. Reminds me of all the pilots crashing small air crafts, because they are so dependent on auto pilot.

One assumes AI could be trained to detect "anomalies" and flag them in results, shortening your workflow since I know when I've seen reports from procedures and tests they can sometimes include hundreds of images that I assume some radiologist needs to troll through but I don't know if I'd trust results that were 100% AI driven. Just like I'm not 100% certain I trust AI "self driven" cars regardless of how great everyone claims the tech is. In the end tech is only as good as its creators...

I wonder what the new Minister of AI will do @EvanLSolomon ?

Dave, in the olden days it might have not made a difference. As the clinician traipsed down to the radiology suite with coffee in hand. Now it depends on the clinician..to accept AI? assuming Rad on site. BTW, how do you sue AI? ..what a mess?

The waiting period for radiology exams is lengthy, the population is aging, which only means we need more radiologists, and up-to-date radiography equipment in our hospitals!

We need professional screening for MRI reqs. The MVA with WAD 1 7 years ago doesn't need a 4th MRI.


