- The chapter rule, up front: AI explains; doctors decide. It is a brilliant explainer and a terrible physician. Never start, stop, or change a medication because an AI said so. Never let it talk you out of (or into) a treatment. And chest pain, stroke signs, trouble breathing, a bad fall — that's 911, not a chat window. Nothing below overrides this box.
Now the good news: used inside those lines, AI fixes the worst part of modern healthcare — nobody explains anything to you. Seven-minute appointments, jargon-stuffed paperwork, test results that read like chemistry exams. The AI has all day, and it never makes you feel slow for asking.
The five things it's genuinely great at
- Translating doctor-speak. "My report says 'mild degenerative changes of the lumbar spine.' What does that mean in plain English?" Thirty seconds later you actually understand your own chart. Follow up with "should I be worried?" and it'll give you the honest, calibrated version.
- Prepping you for appointments. The visit is seven minutes — arrive loaded: "I'm seeing my doctor Tuesday about my knee. Help me make a short list of the questions worth asking." Patients with a list get better appointments. Every doctor will tell you so.
- Explaining (not choosing) medications. "What is metoprolol usually prescribed for, and what should someone taking it generally know?" — great question. "Can I skip my dose?" — that one goes to your pharmacist, who is, by the way, the most underused free expert in America.
- Decoding bills and insurance. This is really Chapter Seven wearing a stethoscope — "explain this Explanation of Benefits," "draft an appeal for this denied claim." Medical billing errors are common, and a calm letter catches them.
- Being there at 2 a.m. Worry doesn't keep office hours. "Is it normal for a bruise to spread like this?" gets a measured answer instead of the horror-story lottery of a search engine — and it will tell you plainly when the right answer is "call your doctor in the morning" or "go in now."
How to ask, step by step
- Describe, don't identify. Say "a 78-year-old who takes a blood thinner" — never your name, Medicare number, or account numbers. The AI gives the same quality answer either way; only the privacy changes. Photos of paperwork: thumb over the numbers, same as Chapter Seven.
- Give it the context that matters: age range, what was diagnosed, what the doctor already said. "My doctor said X and prescribed Y; help me understand the thinking" gets gold. Mystery-symptom guessing games get astrology.
- Ask for the question list, not the verdict. The single best health prompt in this book: "What questions should I ask my doctor about this?" It turns the AI into your briefing aide instead of your physician — which is exactly the right job for it.
- When it's serious, get the human second opinion. AI is a first-opinion machine (Chapter Five). For anything that changes what's in your pill organizer, the chain is: AI explains → you understand → doctor or pharmacist decides.
One more painted line
- Beware "AI doctor" products. Apps, websites, and ads selling AI diagnoses, AI miracle supplements, or "what your doctor won't tell you" — that's Chapter Eight's scam pattern in a lab coat. The real assistants (Chapter Six) don't sell cures, and real medicine doesn't arrive by Facebook ad.
The honest summary: AI won't replace your doctor. It replaces the confusion between doctor visits — and that might be worth more.