Welp. More clinicians are writing about AI, like this one at the New York Times. I like the ambivalence of this doc who both likes using AI tools herself and fears the never-skilling effect on students, residents, and newer clinicians.
Yes we live in interesting times. AI scribes for completing progress notes in real time. Inbasket AI to draft replies to voluminous patient messages. Chart summaries of hundreds of notes and test results. Medical references you can ‘chat’ with to explore anatomy and physiology.
These tools are great for … the last great generation of clinicians trained with years of effortful cognitive friction, like myself and the older clinicians around me. AI assistants are particularly beneficial as an adjunct helper when an experienced clinician needs a nudge and can apply clinical judgement to AI output. ‘That doesn’t sound right’ or ‘oh yes I remember this, good idea.’
This is less possible with new learners who have not put the hard years in to build ‘the knowledge’ (London can drivers amazing internal map of a super complex city and how to drive across it effectively, a task this takes years and demonstrably hupertrophies the parts of the brain involved in navigation). Similarly, clinicians must build ‘the knowledge’ that grants judgement, the ability to distinguish ‘sick or not sick’ from the doorway of a patient’s room.
What to do at this critical crossroad?
No easy answers. Certainly, change our curriculum to create new opportunities for learners to apply productive cognitive friction, if ‘handwriting notes’ and ‘going to the library to seek out the right diagnosis and treatment’ are simply keyboard tapping or ‘chat with my bot’ now.
I for one am not looking forward to my own future infirmity and going to see a clinician who will simply say ‘Hmm. I don’t know. Let’s see what the AI thinks, shall we?’
=shudder-=
Stay in Touch
Get new posts by email — no spam, unsubscribe any time.