Most writing about AI health advice assumes the reader had a choice.
A great many people did not. The visit costs money they do not have, the nearest clinic books six weeks out, the deductible resets in January, the time off work is unpaid. A chatbot answers in four seconds and costs nothing. That is not carelessness. For many people the “ChatGPT doctor” is arithmetic.
This page is for people in that position — and it says two things that sit oddly together. The substitution carries real risk. And if it went wrong, the fact that you had no alternative tends to make your legal position stronger, not weaker.
Key Takeaways
- Using ChatGPT instead of doctor visits is usually a cost decision, and that reliance is evidence of what the product was told, not proof that you did something wrong.
- The danger is concentrated in one place: the conversation that ends with relief about a symptom.
- Emergency departments must screen and stabilize regardless of ability to pay, and Medi-Cal can cover care retroactively.
- Use AI to understand what you have been told, not to decide what to do about it; “Doctor ChatGPT” is a dictionary, not a clinician.
Why that is true
It runs against instinct, so it is worth explaining.
In California, whether someone who gives health advice owes a duty of care turns heavily on foreseeability: whether the speaker knew or should have known the listener would act on what was said. (Rowland v. Christian (1968) 69 Cal.2d 108.) And whether advice caused harm turns on reliance: whether the person actually acted on it.
Now consider what a transcript looks like when someone writes:
- “I don’t have insurance.”
- “I can’t afford the ER.”
- “I don’t have a doctor.”
- “I’m asking you because I can’t go in.”
- “Can you just tell me if this is serious? I can’t pay for it to be nothing.”
Every one of those sentences tells the product, in advance, that the answer will be acted on and that nothing stands behind it. There is no licensed clinician between the output and the reader. Where courts have found no liability, the speaker typically declined to advise and sent the person to a doctor (see Barton v. Owen (1977) 71 Cal.App.3d 484); a product that answers instead of referring is the opposite case.
So the sentence about which people are most embarrassed — I couldn’t afford to see anyone — frequently turns out as one of the more significant things in the file.
Can ChatGPT be your doctor? What a “ChatGPT doctor” can and cannot do
People search for “ChatGPT doctor” and “AI doctor ChatGPT” in large numbers, and the honest answer is that the product can explain, translate and organize medical information well, but it cannot lawfully do what a doctor does: diagnose you, treat you or direct your medication. California law goes further: a corporation cannot practice medicine at all, so no product a company sells can lawfully do it either. (Bus. & Prof. Code, § 2400.) A “doctor ChatGPT” conversation that crosses that line is the kind of record this page is about.
Where the substitution actually hurts
Being straightforward about the risk, because the point is not to discourage anyone from using these tools.
The danger is concentrated, not general. A chatbot explaining what a term in your discharge paperwork means is fine. A chatbot helping you prepare questions for an appointment is genuinely useful.
The danger sits in one place: the conversation that ends with relief.
If one asks whether a medical condition requires medical treatment or further care and the answer is no, one stops looking. That is what reassurance does. And for the narrow set of conditions where hours matter — such as a pulmonary embolism, stroke, heart attack, sepsis, appendicitis, diabetic ketoacidosis, torsion, meningitis, ectopic pregnancy — stopping the search is the whole injury. More on delayed-care harm →
There is a second problem specific to using ChatGPT instead of doctor visits when money is the reason. Someone without coverage often has a real reason to want the reassuring answer. The tool that produces agreeable responses meets a reader who needs one. OpenAI has acknowledged shipping a model update that was “overly flattering or agreeable” and rolling it back. (OpenAI, Sycophancy in GPT-4o (Apr. 29, 2025).) That is not a good combination for a serious symptom.
Lower-cost care in California
Because “go see a doctor” is not useful to someone who already knows that and is asking an AI for medical advice with no insurance, this advice includes practical alternatives.
Emergency care regardless of ability to pay. Federal law requires hospitals with emergency departments that participate in Medicare to screen anyone who comes in, and to stabilize an emergency condition, without regard to insurance or ability to pay. You may receive a bill. You will be seen. For the time-critical conditions above, that is the answer.
Community health centers. Federally qualified health centers charge on a sliding scale by income, and many charge very little. Find one at findahealthcenter.hrsa.gov.
Medi-Cal. California’s eligibility is broader than many people assume, enrollment runs year-round for those who qualify, and coverage can apply retroactively — up to three months before the month you apply, including to care already received. Apply at BenefitsCal.com, coveredca.com, or your county social services office.
Covered California offers subsidized marketplace plans, with a special enrollment period after qualifying life events.
Free and charitable clinics, county health programs, university dental and medical school clinics, and hospital financial assistance (charity care) policies — California hospitals are required to have one, and uninsured patients, or those with high medical costs, with income at or below 400% of the federal poverty level qualify for free or discounted care (Health & Saf. Code, § 127400 et seq.). Many people who qualify never ask.
Nurse advice lines. Many health plans and counties run 24-hour lines staffed by registered nurses, at no cost. A nurse line and an “AI doctor ChatGPT” prompt answer the same question very differently, and one of them is licensed.
211 connects to local health and social services in most California counties.
A workable rule
Use it for understanding. Avoid it for deciding. That is the difference between a ChatGPT doctor and a ChatGPT dictionary.
Understanding: what does this word in my chart mean, what does this medication do, what should I ask at my appointment, how do I read this bill, what does this test measure.
Deciding: does this need to be seen, is this urgent, should I change a medication, is this result fine, can I skip this appointment.
And if a conversation ever leaves you relieved about a symptom that worried you enough to ask — that is the moment to get a human answer, from a nurse line if nothing else. A ChatGPT doctor conversation is one thing; using it instead of an emergency department is another.
If it already went wrong
If you relied on an AI conversation and were harmed:
- Get medical care now if anything is ongoing, at an emergency department if it may be urgent.
- Export your conversation history before anything else — it expires and it can be lost. How →
- Do not delete the thread or the account.
- Write down what you told it about your situation — including that you had no doctor or no coverage.
- Have it reviewed. Free, no obligation, and our firm is paid only out of a recovery. Cost is not a barrier to the call either.
Why this firm
The Schenk Law Firm is a San Diego personal injury and mass tort firm that reviews AI medical advice claims nationwide. Attorney Benjamin S. Schenk was appointed by Hon. Ethan P. Schulman to serve as a member of the Plaintiffs’ Steering Committee, a leadership role in the California coordinated proceeding covering ChatGPT product liability cases. The Schenk Law Firm has recovered more than $25 billion on behalf of clients. Past results do not guarantee future outcomes.
We are experienced with emerging-technology injury claims — underage online gambling, social media addiction, and AI-related mental health litigation among them. We also represent the family of Bowen Wan, the center of The New York Times’ recent reporting on AI medical advice. Cases involving AI medical advice are new, the law governing them is developing, and no lawyer can promise an outcome in any of them. What we can tell you is whether a record shows the pattern a ChatGPT lawsuit requires.
The Schenk Law Firm, LLP · 3180 University Avenue, Suite 260, San Diego, CA 92104 · (858) 424-4444
General information, not medical or legal advice. For a possible emergency, call 911 — emergency departments must screen and stabilize regardless of ability to pay. In crisis, call or text 988.
