Tuesday, May 26, 2026
When AI shows us our real shortcomings
A fascinating article in the New York Times titled "Doctors, This is Why Our Patients are Using ChatGPT," written by Columbia Univ. medical professor Helen Ouyang, lays out an uncomfortable truth: sometimes AI is more humane than humans. She describes her treatment for some problems revealed by blood tests. Her own doctor was barely there for advice, urging "diet and exercise" before signing out. ChatGPT, on the other hand, offered her specific advice, allowed for a dialogue, and accompanied suggestions with encouragement. In other words, the computer creation provided the human touch that actual humans in the medical field often don't offer.
As much as the article offers positives for this one AI application, it also is a harsh critique of how humans are performing those same tasks. Medical care-- driven by profit/cost metrics and other factors-- just don't do much with the "care" part of the job.
In my own field, something similar is playing out. Big law firms are finding that AI can do much of the discovery work that young associates usually handle, and clients are pushing them towards that efficiency. That transition reveals a truth I have known since I was a young associate: that for many of those civil discovery tasks, you never needed someone billing the high rates charged for associate work. Paralegals or others could have done it just as well for less. But... the profit model for law firms depends on billing out lots of associate time.
Overall, I see more downside with AI than upside (and I will say more about that tomorrow when I write about the fascinating encyclical issued today by Pope Leo). But if it does a little to push us towards being more human and caring, I see the good in that. Thanks to Prof. Ouyang for telling that truth.
