On July 23, OpenAI began rolling out Health in ChatGPT, to logged-in U.S. adults. With permission, ChatGPT can draw from Apple Health, supported medical records, One Medical, and Function Health—bringing medications, laboratory results, recent visits, sleep, activity, and other health data into the conversation. OpenAI says more than 300 million people ask ChatGPT health-related questions every week.
This puts to rest the idea that patients are not using AI to help navigate their health. They are and AI can increasingly work from parts of the medical record.
The chatbot may have read part of the record. The more important questions are whether it read the right records, whether they were current, and whether the chart captured the reasoning behind the plan.
Older adults are part of this shift. Pew found that 23% of adults 65 and older have used an AI chatbot, including 19% who have used ChatGPT. Among adults 50 and older who use AI, AARP found that 41% had already asked an AI-powered tool a health-related question. Physicians have moved quickly, too: the AMA reports that more than 80% now use AI in a professional context—about twice the share reported in 2023.
The patient has a copilot. So do we.
Ours summarizes research, drafts the note, and prepares the follow-up. Hers explains the diagnosis and medication list at home, in plain language, as many times as she asks.
So what was she trying to understand after the visit that we had not made clear enough?
The second visit happens at home
AARP asked adults 50 and older who used AI around their healthcare what they did with it:
64% looked up a diagnosis or condition their provider had mentioned.
62% researched a medication that had been discussed or prescribed.
41% used AI to interpret something their provider told them.
45% asked whether a symptom was serious enough to contact their provider.
That last one is triage—even when no clinician knows it happened.
The doctor–patient relationship has not been completely outsourced to AI. But the feedback loop between patients and AI is tightening, while the clinician is often left outside it. The result is a widening care gap between what the patient understands, what AI infers, and what the clinician knows.
Trust has not disappeared. Among older adults who had used AI for health information, 47% said humans provide better guidance, while only 10% thought AI was better. AARP also found that 69% of adults 50 and older were uncomfortable sharing personal health information with an AI tool.
They still want us. They ask the chatbot anyway because it is 9 p.m., the visit is over, and the portal reply may take two days.
AI repeats the explanation without impatience. It decodes terminology, helps a caregiver organize the story, and generates questions for the next appointment. But it works from whatever context it receives.
That is where the clinical risk begins.
The risk is context drift, not just hallucination
A chatbot does not need to invent a fact to lead a patient in the wrong direction. It can reason convincingly from incomplete or outdated information: an old medication list, an isolated laboratory result, or a record that never captured why the plan was chosen.
Access to data is not the same as access to clinical reasoning.
The larger risk is context drift: the chart, the patient’s memory, the caregiver’s summary, and the chatbot history stop telling the same story.
AI flags an interaction, and the patient quietly stops a medication. It interprets one abnormal result without the longitudinal context, and the next visit begins with un-scaring. As one physician on our platform put it: it used to be, “I Googled it.” Now, “Dr. Chat made some conclusions for them.”
We do not yet know how often these conversations change treatment or delay care. But the workflow problem is already here.
Neither side says it out loud
Only 19% of adults 50 and older believe their provider uses AI, and just 9% say a provider has explained how AI was involved in their care. Yet most want to know when it is being used.
Patients rarely disclose their AI use, either.
That matters most for patients with multiple diagnoses, long medication and supplement lists, several specialists, or a caregiver helping coordinate care—the same patients who may benefit most from AI and be most vulnerable to missing context.
What to try this week
Run a two-sided AI reconciliation with five patients:
“Did you use AI before this visit or after your last one?”
“What did you ask it?”
“Did the answer change anything you took, stopped, delayed, or planned to do?”
“Would you like to know when our practice uses AI in your care?”
Chart the answers. If AI changed a medication or care decision, reconcile it like any other unreported clinical change. Nina Vasan and Saneha Borisuth have argued that chatbot use belongs in routine intake. They’re right.
Then send home an after-visit summary that can survive the paste: what you found, why it matters, why you chose the plan, what comes next, and when to call.
Remind patients that consumer AI may not protect health information the same way your patient portal does.
A chatbot may know the patient’s question. You hold the context.
Make your plan the best document she has. Assume it will be pasted into ChatGPT—because it will.
Be a modern clinician with the help of Ultralight, the AI-native EHR built specifically for functional, integrative, and longevity medicine. We’ve recently launched wearables integrations and improved AI-native clinical workflows - get in touch to see the latest updates.
Join us: Functional Medicine Is Longevity Medicine
On September 18 and 19, Dr. Kara Fitzgerald, functional medicine clinician, researcher, and host of New Frontiers in Functional Medicine, runs her annual virtual masterclass for clinicians under this year's theme: functional medicine is longevity medicine. Two days, fully online, built for practicing clinicians. Ultralight will be there.
In the news
Testosterone therapy without documented hypogonadism tracked with 51% higher cardiovascular risk over 10 years. In an eBioMedicine cohort of 358,957 men starting testosterone across 123 health systems, the roughly one third treated off-label showed 51% higher risk of major cardiovascular events and 90% higher all-cause mortality than matched comparators. EHR data with the usual confounding, but this is the study your TRT patients will bring in, and two confirmed morning levels plus a documented indication are what make a protocol defensible against it.
JAMA argues autonomous AI could beat both physicians and physicians using AI by 2030. The viewpoint curates studies since 2024 across five core clinician tasks, from gathering information and diagnosis to chronic disease management, and leans on a meta-analysis of 106 experiments with an uncomfortable finding: when AI already outperformed humans on a task, adding the human back made performance significantly worse. Read it with the counterargument in hand. Performing a clinical task and practicing medicine are different jobs, ownership of the outcome belongs to the second, and the system worth measuring is the physician plus AI in a well-designed workflow, a combination most of these studies barely tested.
AI scribes gave clinicians back 16 minutes of documentation time per eight hours of clinic, plus about $167 a month in added visit revenue. A JAMA analysis of more than 8,500 clinicians across five academic systems, published in April and worth the late look, found adopters spent 13 fewer minutes in the EHR per eight scheduled hours, with heavy users seeing double the EHR savings and triple the documentation savings. The first large multi-site answer to whether the scribe pays for itself.
Upcoming Conferences & Events
Sept 18-19, Dr. Kara Fitzgerald Masterclass “Functional Medicine Is Longevity Medicine”· Online · Led by Kara Fitzgerald, this year’s Masterclass will continue under the theme that functional medicine is longevity medicine. Not as a trend, but as the most science-backed, clinically effective path to extending healthspan. Ultralight will be there!
Sept 22–23, MVMNT Longevity Medicine Summit · Coronado, CA · Evidence-graded longevity science, hands-on labs, and clinical frameworks you can implement the week after. Capped at 300 clinicians. Ultralight will be there!
Oct 8–10, A4M Women's Health Summit · San Antonio, TX · The best clinical education on hormone, metabolic, and midlife women's health you will see this year. The room to be in if you are growing the perimenopause and menopause side of your practice.
Oct 17-18, Roundtable of Longevity Clinics · Buck Institute, Novato, CA · Some 250 longevity clinic leaders, physicians, and researchers working toward shared standards for longevity testing and interventions. In-person and virtual tickets are open. Ultralight will be there!
Oct 21-23, DOC (Living Room Lab) · Sonoma, CA · Salon-style sessions on longevity science and medical AI with faculty from UCSF, Stanford, and the Buck, plus validated diagnostics in the Living Room Lab. Ultralight will be there!
Oct 21–24, NAMS Annual Meeting · San Diego, CA · The single most practice-changing meeting of the year for midlife women's health. Your protocols will look different after this one.
Nov 5–8, Eudēmonia Summit · West Palm Beach, FL · One of the most talked-about longevity gatherings in the U.S. Experientials, hands-on demos, and the best place to try the emerging frameworks your patients will ask you about next year. OvationLab and Ultralight will be there!
Nov 5-7, Private Physicians Alliance Annual Meeting · St. Petersburg, FL · The gathering for independent, cash-pay, and concierge physicians navigating practice independence. Practical and peer-driven. Ultralight will be there!
Nov 8-11, American College of Lifestyle Medicine Conference · Orlando, FL · Lifestyle medicine's main annual event — evidence-based approaches to behavior change, chronic disease, and healthspan. Growing overlap with the longevity medicine community.
Dec 11–13, A4M Longevity Fest · Las Vegas, NV · The biggest longevity event in the U.S. The room spans clinicians, industry, founders, and the people building next year's platforms, and the connections from this one tend to compound through the rest of your year. OvationLab and Ultralight will be there!
Know of an event we should add? Reply and tell us.
Until next week
Both sides of the visit have a copilot now, and they have not met. The patient's copilot has her question. Ours has the chart. Asking what she asked, and sending home a plan that survives being pasted into ChatGPT, keeps the two stories one story.
Reply and tell us the most surprising AI answer an older patient has brought you. The best ideas in this newsletter come from clinicians doing the work.
Until next week, keep building the practice you imagined when you started.
— Dr. G and Sunita