Here's a Thought: Is Paging Dr. AI a Good Idea?

In today's culture, we crowdsource everything from brand names to bread starters. Here's what to know about handing your healthcare over to an algorithm.

✍️ By Jenna McCarthy

---

I’m a former beauty editor and unapologetic product junkie with historically bulletproof skin. I get a pimple every five years—maybe—and I’ve never met a Vitamin C serum, retinol, or crushed-apricot-exfoliating-miracle I didn’t enthusiastically overdo. So I was more than a little surprised to wake up one day with a strange forehead rash. It was nothing dramatic—a little redness, a few bumps—and surely something I could cure with my arsenal of natural remedies.

After two weeks of slathering my skin with hourly rounds of coconut oil, tallow, and some tea tree-aloe-castor oil concoction I discovered on TikTok, the rash wasn’t just worse—it was spreading. I now had angry red blisters across the bridge of my nose, too.

A normal person might go see a dermatologist at this point. I am not a normal person.

Instead, I did what any modern, health-conscious, mildly distrustful individual does: I Googled. I fell into a rabbit hole of forums, Reddit threads, and holistic wellness blogs written by women named Sage who live in Sedona and have very strong opinions about deodorant.

If it had ever been described as “soothing,” “healing,” or “ancestral,” it was in my cart and on my face. I bought creams sourced from sea creatures and balms crafted from arctic Norwegian plants. Strangers in chat rooms boasted about turning their irritated skin into a Snapchat filter overnight with evening primrose, geranium, rosemary, and something called Helichrysum oil, so I bought them all and layered one on top of the other. I tried homeopathic sulfur soap and organic diaper balm (I kid you not) and I hope you’re sitting down, because all of my efforts were making the situation progressively worse.

The rash was now dancing across my cheeks.

Reluctantly, I called my dermatologist, who very cheerfully informed me they’d be delighted to see me… in late October. Desperate, I uploaded photos of my hideous, furiously red face to ChatGPT.

“Jenna,” Chat replied. “Respectfully… what have you done?”

Within seconds, I had a diagnosis. It was not dramatic. I did not have a rare condition or an exotic deficiency. It was essentially: you overdid it with the products—yes, even the “natural” ones—and basically destroyed your skin’s protective barrier.

And then came the “treatment plan.” Not a cream. Not a serum. Not a 12-step routine involving ethically sourced sea moss.

Nothing.

Literally, that was the plan. Do nothing. Stop all products. No soap, no moisturizer, no oils, no balms, no magical potions. Let your skin calm down. Leave it alone—like a Victorian child recovering from the vapors. And here’s the part that floored me: It worked.

In less than two weeks, the rash was gone. No prescription. No follow-up appointment. No “you’ll need to stay on this indefinitely.” No discount code for my next order or helpful suggestion that I might also benefit from a matching toner. Just: stop.

I’ll be honest. I felt guilty about “doing my own research”—thank you, Covid—but that faded pretty quickly when I realized the alternative was waiting five months to pay $250 for 13 minutes with a dermatologist and an Rx my insurance wouldn’t cover—not to mention whatever fresh horrors I’d have unleashed on my face in the meantime.

Apparently, I’m a little slow on the uptake. Hundreds of millions of people are now using AI for health questions every week, and not just out of curiosity. According to a recent Ohio State survey, 62% use it to understand symptoms before deciding whether to seek care, and 44% consult their preferred bot to make sense of diagnoses or test results. Not necessarily instead of doctors, mind you—but before them, alongside them, and sometimes to figure out what to ask when they finally get there. (That part is actually key.)

Meanwhile, the people actually in medicine aren’t exactly ignoring the technology. Sixty-six percent of physicians surveyed in 2024 reported using AI in their practice (a 78% jump from the prior year—so surely that number is far larger now). AI is particularly useful in enhancing diagnostic accuracy, with studies showing a 20% increase in breast cancer detection, far fewer false positive tests, and significant improvements in identifying conditions like food allergies. In other words, “Paging Dr. AI” isn’t a joke anymore. It’s the world we live in.

Which makes sense, because machine learning offers something modern healthcare can’t: unlimited time. Time to take in your entire health history. Time to ask follow-up questions. Time to sit with you while you say, “Okay, but I tried that already, and what if it’s this?” five different ways. It feels like having a dedicated 24-hour medical team that never glances at the clock.

Will people misuse it? Of course! They already misuse Google. They already self-diagnose. They already ignore medical advice in favor of something their neighbor’s Pilates instructor swears by. But this feels like an upgrade. Because instead of doing nothing—which is what lots of people do, sadly—you’re at least doing something. And we’ve gone from “I have a headache, I’m probably dying” to “Here are three plausible explanations, none of which involve immediate death.” Plus, unlike your group chat, AI actually has some guardrails and isn’t shy about telling you when it’s time to see a professional. I’ll take that.

There’s a catch, of course, and it’s a big one: AI is only as good as what you tell it. Feed it half the story, leave out a medication, forget that one weird allergy, and congratulations, you’ve just crowdsourced your own misdiagnosis. Dr. AI will confidently connect the dots you hand it, even if you hand it the wrong dots. Which is how you end up feeling either completely reassured about something that isn’t fine—or newly convinced you’re harboring a rare condition that requires immediate attention and possibly a specialist in Zurich.

But the even bigger catch, the one that doesn’t get talked about enough, is that AI is also only as good as what it was trained on. And what it was trained on is the official narrative—peer-reviewed literature, institutional consensus, pre-approved guidelines. If you believe—as I do—that we’re not being told the full story about certain medications, certain vaccines, certain public health decisions made at warp speed under emergency authorization, you can bet that your AI assistant does not share your skepticism. It was trained on the press-release version of reality and it will reflect that press-release version back to you, with great confidence, every single time.

There’s also the little matter of privacy. Before you describe your most intimate symptoms to an AI, know this: unlike your doctor, most AI platforms aren’t bound by HIPAA, which means your detailed medical queries could be stored, reviewed by human trainers, swept up in a data breach, or quietly shared with data brokers—the internet’s shadowy middlemen who exist solely to know things about you and sell them to advertisers. So no, your boss probably won’t find out. But somewhere, a targeted ad algorithm is about to become extremely invested in your recovery.

To be completely fair, my issue was a skin rash. A little unsightly? Absolutely. A life-or-death medical emergency? Highly unlikely. And in my case, the mainstream answer happened to be the right one—“stop putting things on your face” is not exactly a controversial position in dermatology. If I found a lump in my neck or started spontaneously bleeding from my ear, I promise you, ChatGPT would not be my plan A. (Fine, it might be my plan A… but it would not be my final destination.) When navigating your own healthcare, the key is being savvy enough to know when to ask a computer a question and when to default to a living, breathing human with years of education, training, and actual real-world experience.

Despite its success in “treating” my rash, I do not believe AI is here to replace doctors. It’s here to make patients a little less confused, a little more informed, and a lot less likely to panic-buy twelve different creams and accidentally declare war on their own face. But when it comes to the big stuff, there is still no substitute for a trained specialist who can look at you, listen to you, notice what you’re not saying, and ask the one question you didn’t think to include in your prompt.

The bottom line: AI can process information, but doctors are trained to put it into context. One helps you ask better questions—the other makes sure you get the right answers. Proceed wisely.

Do you consult AI with medical questions and concerns? Let me know in the comments!