The baggage claim problem of AI in medicine
- Zohaib Akhtar MD MPH
- 23 hours ago
- 5 min read
You know, one of the oddest things about modern life is how we all insist we have no time. We say it constantly: I don’t have time to exercise. I don’t have time to cook. I don’t have time to just sit and think. And yet somehow, we find three hours to binge-watch Netflix or two hours to scroll on our phones.
So it’s not that the time isn’t there. It’s that the time doesn’t feel like it belongs to us. That’s the cultural trick, especially in the United States. Every moment gets chopped into fragments (emails, meetings, errands, commutes) until the whole day feels like a race. Because speed has become the measure of progress, anything slow feels like failure.
You can see it in how we consume information. Nobody claims to have time to read, but we all scroll. Endlessly. Skimming, swiping, chasing the next thing without even knowing what we’re looking for. Our attention span has been sliced into fragments, measured in taps and swipes instead of minutes and hours.
And here’s the irony: speed feels good in the moment. Amazon Prime, fast food, same-day delivery...... We even need coffee, the drug that makes everything feel rushed and makes us feel of squeezing more into the same crowded day. But humans don’t actually work like that. We don’t measure time in stopwatch seconds. We measure it in how it feels: Did I feel listened to? Did I feel rushed? Did I feel cared for?
Airports learned this the hard way. For years, passengers complained about baggage delays. So airports invested millions in faster belts and more automation. Complaints barely changed. Then one airport tried a different idea: don’t speed up the bags, slow down the passengers. They moved the gates further from the carousel. By the time people arrived, their luggage was waiting. Complaints collapsed. Nothing about delivery speed changed, only the experience of time. Its because we measure time in experience, in meaning, in how it feels.
Now let’s apply this to Healthcare.
What burns doctors out? Not the time with patients. Nobody says, “I wish I could see fewer patients.” What burns us out is the time stolen from patients. The twenty minutes you could spend listening to the real story, puzzling through a symptom, or just sitting in silence while someone gathers the courage to say what’s really wrong -- instead wasted on billing codes and EMRs. That’s the baggage carousel moment in medicine: time wasted in the worst possible way.
And patients feel it. They don’t complain if results take a couple of days. They complain when results pop into a portal instantly, stripped of context, and terrify them. They don’t complain about a twenty-minute visit. They complain when half of it is spent with the doctor staring at a screen. Every clinician knows the moment: the real truth only emerges as the hand reaches for the door handle. That’s not inefficiency. That’s medicine.
Different systems shape this differently. In the U.S., every clinical minute carries a dollar sign. A visit isn’t just a block of care, it’s a coded, billable unit that can run into hundreds of dollars. Empathy feel expensive. In South Asia, it’s harsher still: too few doctors, impossible patient loads, empathy squeezed out by necessity. And then in Scandinavia, I saw something different. Care felt slower. Not always longer, but richer in texture. Empathy wasn’t squeezed in -- it felt built in.
It’s not that Scandinavian doctors are saints while American doctors are machines. It’s that the scaffolding is different: smaller populations, publicly funded hospitals, and a system that protects the space for presence.
So no, it’s not a fair apples-to-apples comparison. But this is exactly where AI matters. Because its a chance to reconcile the tension: to keep the business viable without asking empathy to die.
Here’s the irony. Clinicians love medicine. Doctors, nurses, PA's endure years of training. They love the diagnostic dance, the conversation, the satisfaction of seeing someone walk out better than they came in. That’s the joy. That’s the journey. And yet we keep building AI tools that optimize the wrong thing --- faster notes, faster throughput, more patients per hour. We’re not making the consultation richer. We’re just making the baggage belt spin faster.
Other industries already learned that slowness has value. Starbucks, for example, instructed baristas to limit themselves to making two drinks at once to reminded customers that each beverage is created with intention and care, easing the premium price .Fine dining works the same way: a dish that arrives too quickly feels cheap, rushed, like fast food. The pace itself signals value. Now imagine medicine reversed: patients come expecting crafted care, and we hand them a pre-packaged latte from the fridge, all in the name of “efficiency.”
So what should AI do? Not replace that time. Protect it. AI should quietly handle the notes, the billing, the inbox flood -- so the doctor can actually look at the patient. That wouldn’t make medicine faster. It would make medicine slower where it matters. Not a chatbot firing answers. Not a dashboard spitting out risk scores before the exam is finished. Slow AI would track health over months, surface patterns gently, and give clinicians back the freedom to be present. It would guard empathy instead of crowding it out.
But the risk is obvious. Give this to an administrator, and they’ll say: “Wonderful! Now each doctor can see 40 patients instead of 20.” That sounds like progress --- until you realize you’ve optimized for throughput, not healing. It’s like saying: “We’ve improved the theater by doubling the plays each night.” Wonderful.... except you’ve destroyed the performance.
I’ve seen medicine at its best — slow when it matters, empathetic even under pressure. And I’ve seen it at its worst — burned out, cynical, stripped of joy. Some days I’ve been closer to one version, some days the other. And that’s why I get nervous about AI. Because if it’s designed only to make us faster, it will lock in the worst version of ourselves: the rushed, distracted, half-present doctor.
What should we do?
For AI developers: Stop assuming faster is better. Design tools that return time to the clinician–patient relationship, not strip it away.
For Hospital administrators: Measure the right outcomes. Patient experience and clinician well-being are not inefficiencies, they’re leading indicators of quality. My group demonstrated this by introducing an efficient clinic at Johns Hopkins Medicine that optimize care with improved quality.
For Clinicians: Demand “slow AI.” Ask not how it speeds you up, but how it protects your presence.
Because faster isn’t always better. And if AI — and we — can’t tell the difference between when to be fast and when to be slow, then it may end up being a cure for bureaucracy, but a disease for care.
((Z))