Leadership1 publisher2 min readPublished
Medicine holds the public's one favourable view of AI while hospitals deploy it on notes
Seventy percent of American adults say they are more worried than excited about AI, and the one exception in the polling is scientific research and medicine. The fastest uptake inside medicine is administrative.
The Board Room · Leadership desk
What happened
- A recent NBC News/Decision Desk poll found 70 percent of American adults say they are personally more worried than excited about artificial intelligence, according to Vox's account of the survey.
- Vox reports that opinion on AI is underwater in every area of society the poll covered, with its use in scientific research and medicine the one notable exception.
- Dhruv Khullar, a Weill Cornell Medicine physician who uses the tools in his own practice, said healthcare is taking up AI faster than any new technology he has seen.
- The most rapid uptake, by Khullar's account, is administrative: AI scribes now take notes during visits, and he says patients seem to like it so far.
- Khullar counts drug discovery as the biggest chunk of the favourable public view and says AI is already making a huge dent in the early parts of it.
Compiled by The Board RoomSomething wrong?How this is made
Why it matters
- contradiction The polling exception covers research and medicine, while the surface with the most patient contact inside medicine is administrative and navigational. A health system that cites the poll for a patient-facing tool is citing approval of something else.
- decision A health system choosing between a scribe and a patient-facing navigator is choosing whose trust it needs: the clinician's, or that of a patient who has to act on the answer.
- exposure If adoption is running on frustration with the existing system, the first visible navigation failure is judged by a public that was already majority-worried.
- constraint Any deployment outside research and medicine is recruiting early support from a pool no larger than 30 points.
Set against that 70 percent, the entire remainder of the sample is 30 points wide [1][13]. Everyone more excited than worried, everyone neutral and everyone undecided sits inside it [13]. Vox reports the headline figure and the single exception, and the domain-by-domain numbers stay out of the account [14]. So a health system is left guessing how much of the goodwill belongs to a protein-structure model and how much to a chatbot in a patient portal.
The exception is scientific research and medicine [2]. Research has the cleaner public record of the two: the 2024 Nobel Prize in Chemistry went to researchers who developed an AI model that enabled breakthroughs in predicting the complex structures of proteins [11].
Khullar traced part of the adoption speed to the state of the system he works in. "Part of that might be that the healthcare system is so messed up that there's an appetite for some type of change," he said [6]. If the appetite comes from frustration with what exists, each appointment is a fresh test of whether the new experience beats the old one.
The uses Khullar describes ask for different amounts of patient trust. A scribe works in a room where the clinician is still the one answering, and he said doctors can now look patients in the eye instead of poring over a computer and typing [7]. Patient navigation asks for something different: a person diagnosed with cancer or heart failure acting on what the software says about when the next scan is, or whether a medication needs an empty stomach [8]. The second-opinion use stays inside the profession, and Khullar said he uses it every day in the hospital as clinical decision support [10].
Khullar put a limit on the discovery story himself. "I don't want to leave this conversation and say that we've solved the cure for cancer because of AI," he said [12]. The conversation also set that optimism beside warnings about the technology's risks [15].
So the decade-scale case sits on early-stage drug discovery, and this quarter's choice for a health system is about scribes and navigation, judged by the same 70 percent who told the pollster they were more worried than excited [1]. Khullar's own summary of sentiment carried a time limit. He said that "right now, at least, there's a generally optimistic view of how AI will change healthcare going forward" [4].
What to watch
- Whether the full NBC News/Decision Desk tables are published with domain-level splits, which would show how much of the medical exception covers patient-facing tools.
- Whether the reported patient approval of AI scribes survives the move from note-taking into patient navigation, where the software answers the patient directly.
- Whether AI's contribution to drug discovery moves past the early stages Khullar describes into candidates that reach trials.