Updated Oct. 26, 2026: rewritten as a practical guide for patients on how AI is used in U.S. doctors’ offices and hospitals, your rights and the questions to ask.
Quick answer: AI is already part of most U.S. medical care. An American Medical Association survey released in 2026 found that about 8 in 10 physicians use AI in their practice, up from 38% in 2023. The most common tools are AI scribes that listen to your visit and draft the doctor’s notes, tools that draft replies to patient messages, and FDA-cleared imaging software that flags problems on scans. A licensed clinician is supposed to review AI output. You have the right to ask whether AI is used in your care, and in some states, like California and Texas, providers must tell you.
Here’s where AI shows up, what it can and can’t do, how your data is protected and what to ask at your next appointment.
Where you’ll meet AI in health care
| Where | What the AI does | What to ask |
|---|---|---|
| Exam room | An “ambient” AI scribe records the conversation and drafts the visit note | “Is this visit being recorded? Can I opt out?” |
| Patient portal | AI drafts replies to your messages for a nurse or doctor to edit | “Did a clinician review this answer?” |
| Radiology | FDA-cleared software highlights possible problems on X-rays, CT scans and mammograms | “Was AI used to read my scan, and did a radiologist check it?” |
| Hospital wards | Early-warning tools watch vital signs for signs of sepsis or decline | “What happens if the system flags me?” |
| Billing office | AI suggests billing codes and handles prior authorization paperwork | “Can I get an itemized bill?” |
| Insurance | Insurers use algorithms to review some claims and authorizations | “Was a human involved in denying this? How do I appeal?” |
AI scribes: the tool you’re most likely to see
If your doctor asks, “Is it OK if I use a tool that takes notes for me?”, that’s an AI scribe. It listens to the conversation on a phone or computer, then writes a draft of the visit note. The doctor reviews and signs it.
Doctors like these tools because they cut typing time. The AMA has reported on physicians who say an AI scribe saves them about an hour at the keyboard each day. That can mean more eye contact and less screen time during your visit.
The trade-offs are real, though:
- Mistakes happen. AI can mishear a drug name or add something that wasn’t said. Your doctor is supposed to catch that, but not every error gets caught.
- Privacy concerns. Your visit is recorded, even if briefly. Ask how long recordings are kept.
- You can say no. Most practices let you decline. Your care shouldn’t change if you do.
AI in scans and diagnosis
The FDA keeps a public list of AI-enabled medical devices it has authorized. It now includes well over 1,000 products, and most are used in radiology. They help spot things like lung nodules, brain bleeds and signs of breast cancer.
These tools work as a second set of eyes. A radiologist still reads your scan and signs the report. AI can speed things up, especially when a hospital is busy, but it doesn’t replace the doctor’s judgment.
What the law says about AI and your care
| Rule | What it means for you |
|---|---|
| HIPAA | Your doctor’s AI vendors must sign agreements to protect your health information, just like other contractors |
| California AB 3030 (since Jan. 1, 2025) | If AI writes a clinical message to you and no clinician reviews it, the message must say so and tell you how to reach a human |
| Texas (since Jan. 1, 2026) | Providers must tell patients when AI is used in their care, and doctors must review AI-generated records |
| FDA | Reviews AI tools that diagnose or treat; many note-taking tools fall outside FDA review |
Other states are writing their own rules, so expect more disclosure requirements in 2027.
Also read: How to Negotiate a Medical Bill
6 questions to ask at your next appointment
- Do you use AI in my care? It’s a fair question, and most offices will tell you.
- Is our conversation being recorded? If yes, ask how long it’s stored and who can access it.
- Can I opt out? You can usually decline an AI scribe without any change to your care.
- Who reviews the AI’s work? The answer should be a licensed clinician.
- Can I read my visit notes? Federal rules give you access to most of your notes in the patient portal. Read them and report errors.
- If my insurer denies a claim, was AI involved? You have a right to appeal any denial, and you can ask for a human review.
Made-up examples (not real people): These are made-up examples to show how this works in real life.
Gloria, 67, in Phoenix notices her visit note says she takes a blood thinner she has never used. An AI scribe likely misheard. She messages her doctor through the portal, the note is corrected the same week, and her pharmacy record stays accurate.
Kevin, 41, in Ohio gets a portal reply that reads oddly formal. At the bottom it says the message was drafted with AI and reviewed by a nurse. He replies with a follow-up question and gets a call from the nurse the next day.
How to check whether a health AI tool is FDA-cleared
- Get the product name. Ask your doctor or hospital which AI tool they use, or check the app’s “About” page.
- Search the FDA’s AI device list. The FDA publishes a free, searchable list of AI-enabled medical devices it has authorized.
- Look at the intended use. Clearance covers a specific job, like flagging a possible brain bleed on a CT scan, not everything the tool might do.
- Remember what’s not on the list. Note-taking tools, scheduling bots and general chatbots usually aren’t reviewed by the FDA at all.
What AI still can’t do well
- Physical exams. AI can’t feel a lump, listen to your lungs or notice how you walk into the room.
- Rare conditions. Tools trained on common cases can miss unusual ones.
- Your full story. Your job, family history and what you’re worried about often don’t make it into the data.
- Accountability. If something goes wrong, the licensed clinician and the health system are still responsible, which is why a human must sign off.
Using AI chatbots for health questions
Many people now ask chatbots about symptoms before calling a doctor. That can help you prepare, but be careful:
- Chatbots can be wrong and sound confident anyway. Treat answers as a starting point, not a diagnosis.
- HIPAA usually doesn’t cover them. When you type your own health details into a consumer app, those privacy rules generally don’t apply.
- Don’t share full names, birth dates or ID numbers in a chatbot.
- Bring your questions to a real visit. A good use is making a list of what to ask your doctor.
- In an emergency, skip the chatbot. Call 911 for chest pain, stroke signs or trouble breathing.
Common myths
- “AI is replacing my doctor.” No. In U.S. medicine today, AI assists and a licensed person makes the call.
- “Everything AI does is FDA-approved.” No. Many scribe and paperwork tools aren’t reviewed by the FDA at all.
- “I can’t refuse.” You can usually decline AI scribes and ask for human review of messages and claims.
- “AI makes care cheaper for me.” Not yet, in most cases. Savings to hospitals don’t always reach your bill.
What to watch in 2027
Expect more states to require AI disclosure, more hospitals to roll out AI scribes, and closer federal attention to insurers’ use of algorithms in claim denials. If you’re on Medicare, watch for changes in how plans use automated reviews for prior authorization.
FAQ
Do doctors have to tell you if they use AI?
It depends on your state. California requires a disclaimer on AI-written clinical messages that a clinician hasn’t reviewed, and Texas requires providers to disclose AI use in care. Everywhere, you can ask.
Can I refuse an AI scribe during my appointment?
Usually, yes. Most practices let you decline the recording, and your care shouldn’t change if you do.
Is my health data safe with AI tools?
Tools your doctor uses are covered by HIPAA through agreements with the vendor. Consumer chatbots you use on your own generally aren’t covered by HIPAA.
How many doctors use AI?
An AMA survey released in 2026 found about 8 in 10 physicians use AI in their practice, up from 38% in 2023.
Is AI used to read X-rays and mammograms?
Yes. Many FDA-authorized AI tools help radiologists spot problems on scans. A radiologist still reviews and signs the report.
Bottom line
AI is now a normal part of U.S. health care, mostly behind the scenes: writing notes, drafting messages and double-checking scans. It can give your doctor more time with you, but it can also make mistakes. Ask whether AI is used, read your visit notes, and speak up if something looks wrong.
Last reviewed: Oct. 26, 2026. This article is general information, not medical or legal advice.
Sources: AMA: AI scribes; FDA: AI-enabled medical devices list; ArentFox Schiff: California AB 3030; Becker’s: states regulating AI in health care; HHS: HIPAA business associates.
Featured image: A doctor takes notes on a computer during a consultation. Photo: Shixart1985 / Wikimedia Commons, CC BY 2.0; edited by Nexgen Headlines.
