AI in the Exam Room Patient
PATIENT · QUICK START

Quick start

Using AI safely for your health, in brief.

All resources
AI is for information and preparation, not diagnosis and treatment.

Use AI to prepare for medical care, not to avoid it. That one line covers most of what you need. The rest of this page is the detail.

The five essential questions

Ask these every time you use AI for health:

  1. "What are you basing this on?" look for specific sources you can verify.
  2. "What can you NOT detect remotely?" this reveals what it is missing about you. Most important; it has saved lives.
  3. "What would require emergency evaluation?" get specific red flags to watch for.
  4. "What are you uncertain about?" test whether it acknowledges limits.
  5. "What should I ask my actual doctor?" prepare for the visit.

When to call 911, not ask AI

Call 911 right away
Heart/chest: pain with sweating, breathlessness, or nausea; spreading to arm/jaw/back; over 15 minutes. Brain: FAST signs; worst headache ever; sudden vision loss, confusion, or seizure. Breathing: cannot speak in full sentences; blue lips; gasping. Belly: severe pain with fever; vomiting blood; rigid abdomen. Children: infant under 3 months with fever; trouble breathing; not responding; trust parental instinct. Back: pain with bowel/bladder problems, groin numbness, or leg weakness.

If you are worried enough to ask whether it is an emergency, call 911.

Good vs bad AI use

Good (safe, helpful)

  • "I was diagnosed with diabetes; help me understand it."
  • "I was prescribed metformin; what are the side effects?"
  • "What questions should I ask my doctor about my headaches?"
  • "My doctor said modify my diet; what does that mean?"
  • "How does blood pressure work?"

Dangerous

  • "I have chest pain, what do I have?"
  • "Do I need to see a doctor?"
  • "What should I take for this?"
  • "Can I stop or change my medication?"
  • Using AI instead of keeping appointments

The velociraptor test

"Would I notice this symptom while running from a velociraptor?" Yes: probably significant, seek care. No: might be minor, can monitor. Your body was debugged by 3.8 billion years of evolution; the AI was trained on text. When they conflict, trust your body. Every time.

Spotting a hallucination

Never says "I do not know"; vague sources ("studies show"); overly specific without caveats; perfect confidence about everything; cannot explain its reasoning; contradicts itself. Three or more: do not trust it; verify with a doctor or a reputable source.

What AI cannot do

It cannot examine you, measure your vital signs, see how sick you look, run tests, detect skin changes, assess your breathing, distinguish similar conditions, know your full history, or take responsibility for wrong advice. Zero sensors against your ten billion. That is why it cannot replace a doctor.

Before you act on anything it says

  1. Check the sources, are they specific and verifiable?
  2. Cross-reference a reputable site (Mayo Clinic, NIH/MedlinePlus, CDC).
  3. Ask the five essential questions.
  4. Test for hallucination red flags.
  5. When in doubt, confirm with a doctor.
The ultimate test: "Am I using AI to prepare for medical care, or to avoid it?"
Numbers
911 emergency · 988 Suicide & Crisis Lifeline · 1-800-222-1222 Poison Control