You project a worthwhile objective to develop over the near term. What about 10-15 years from now:
Suppose you need a surgical procedure (say 15 years from now), perhaps brain surgery. You now have a choice: do you want to be operated on by a human surgeon or a surgical robot?
Since the robot is networked into 1000+ other surgical robots around the world, it has their collective experience. Also, the robot has specialized limbs for both sensing and operating, and has fast reflexes. Lots of unknowns about the human surgeon, like did she get enough sleep last night, did she argue with her spouse or kids this morning, is she feeling a little depressed, etc. Were her previous surgeries acceptable, but not optimal.
The challenging truth is that many physicians get stuck on their initial diagnosis and won't let go, even when there is evidence to the contrary. This is a powerful and sometimes dangerous bias. Instead of approaching it as, "AI is highly likely to be right, let me see if there are any reasons it might be wrong in this case," doctors tend to say, "I think it's X and the AI is wrong."
1. From years of experience on many fronts, physician behaviors, (both as doctors and as human beings) MUST be accounted for. Think Durkheim!!!!
2. AI should be thought of as Augmented Intelligence, not Artificial Intelligence. Would help you get unstuck.
3. What you argue for is better thought of using Bloom’s Educational Taxonomy for which “AI today” really exists on the bottom two levels of a pyramid of 6 levels in height.
4. Read Patel’s recent book: Intelligent Systems in Medicine and Health – The Role of AI. The book will impress upon you the EQUAL role for human cognition in incorporating any computer generated anything into clinical practice.
5. Finally, not only do computer tasks need to be married correctly with physician/nursing tasks, but the type of marriage also depends upon the clinical venue, (E.g., Radiology Department versus Emergency Department).
Much to do.
Again, the article is great but only scratching the surface.
I am so glad you posted this on Substack, as the headline on the New York Times article put me off (I could not see who authored it, so didn’t look further). Your point, “The challenge ahead isn't just technological – it's about reimagining the role of both physician and A.I. in healthcare delivery . . . . This may require us to let go of some preconceptions and embrace new models that might initially feel counterintuitive but ultimately lead to better patient outcomes.”
Your analysis makes clear that AI, if employed properly, could work to free overworked, time-pressed physicians to have more time to think and truly attend to the patient in front of them, which we all sorely need.
Your analysis is also applicable to a myriad of other challenges we face right now: we urgently have to stop being reactive, think more subtly, and give ourselves room and time to make the best possible use of the circumstances with which we are presented, however counterintuitive the available options may seem at first.
You project a worthwhile objective to develop over the near term. What about 10-15 years from now:
Suppose you need a surgical procedure (say 15 years from now), perhaps brain surgery. You now have a choice: do you want to be operated on by a human surgeon or a surgical robot?
Since the robot is networked into 1000+ other surgical robots around the world, it has their collective experience. Also, the robot has specialized limbs for both sensing and operating, and has fast reflexes. Lots of unknowns about the human surgeon, like did she get enough sleep last night, did she argue with her spouse or kids this morning, is she feeling a little depressed, etc. Were her previous surgeries acceptable, but not optimal.
The challenging truth is that many physicians get stuck on their initial diagnosis and won't let go, even when there is evidence to the contrary. This is a powerful and sometimes dangerous bias. Instead of approaching it as, "AI is highly likely to be right, let me see if there are any reasons it might be wrong in this case," doctors tend to say, "I think it's X and the AI is wrong."
exactly. Lots of confirmation bias here
Hi Eric,
Just finished reading your NYT op-ed.
Great article, better idea.
Four things:
1. From years of experience on many fronts, physician behaviors, (both as doctors and as human beings) MUST be accounted for. Think Durkheim!!!!
2. AI should be thought of as Augmented Intelligence, not Artificial Intelligence. Would help you get unstuck.
3. What you argue for is better thought of using Bloom’s Educational Taxonomy for which “AI today” really exists on the bottom two levels of a pyramid of 6 levels in height.
4. Read Patel’s recent book: Intelligent Systems in Medicine and Health – The Role of AI. The book will impress upon you the EQUAL role for human cognition in incorporating any computer generated anything into clinical practice.
5. Finally, not only do computer tasks need to be married correctly with physician/nursing tasks, but the type of marriage also depends upon the clinical venue, (E.g., Radiology Department versus Emergency Department).
Much to do.
Again, the article is great but only scratching the surface.
Eric
thanks, Eric. we can only do so much with 900 words. This is all a work in progress
I am so glad you posted this on Substack, as the headline on the New York Times article put me off (I could not see who authored it, so didn’t look further). Your point, “The challenge ahead isn't just technological – it's about reimagining the role of both physician and A.I. in healthcare delivery . . . . This may require us to let go of some preconceptions and embrace new models that might initially feel counterintuitive but ultimately lead to better patient outcomes.”
Your analysis makes clear that AI, if employed properly, could work to free overworked, time-pressed physicians to have more time to think and truly attend to the patient in front of them, which we all sorely need.
Your analysis is also applicable to a myriad of other challenges we face right now: we urgently have to stop being reactive, think more subtly, and give ourselves room and time to make the best possible use of the circumstances with which we are presented, however counterintuitive the available options may seem at first.
Critically important, and I will restack.