More and more older adults are asking AI assistants about their health, from “What does this word on my test results mean?” to “What should I ask my cardiologist?” Used well, AI can make you a better-prepared patient. Used badly, it can give you wrong answers that sound completely convincing.

Here’s how to get the good without the risks.

Plain talk: AI assistant

An AI assistant, such as ChatGPT, Gemini, Copilot or Claude, is a program you can talk to in everyday words. It writes answers based on patterns in huge amounts of text. It doesn’t know you, can’t examine you and sometimes makes things up.

What AI is good at

  • Explaining medical words in plain English: “What does atrial fibrillation mean, in simple terms?”
  • Preparing for appointments: “I’m seeing my doctor about knee pain. Help me write a short list of questions to ask.”
  • Understanding letters once you’ve removed personal details: “Explain this paragraph from a hospital letter in plain English.”
  • Organizing information: “Turn this list of my medicines and times into a simple daily chart.” (Then check it with your pharmacist.)
  • Everyday wellbeing: gentle chair exercises, meal ideas for a diet your doctor gave you, tips for sleeping better.

Type or say this

“I have an appointment with my doctor next week about [your concern]. Help me write 5 short questions to ask, in plain English, in large simple bullet points.”

What AI should never do for you

  • Diagnose you. It can’t examine you, run tests or see your full history.
  • Tell you to start, stop or change a medication. Only your doctor or pharmacist should do that.
  • Decide whether you need care. If you’re worried, call your doctor or a nurse line.
  • Replace an emergency call. For chest pain, trouble breathing, signs of a stroke or a bad fall, call 911 (US and Canada), 999 (UK) or 000 (Australia) straight away.

Safety tip

AI assistants can be confidently wrong. An answer can sound expert and still contain mistakes. Treat it like advice from a well-read friend: useful for getting started, never the final word.

Protect your privacy

  1. Leave out anything that identifies you: your name, date of birth, address, and your Medicare, Social Security or health insurance numbers.
  2. Describe, don’t identify. “A 78-year-old man with type 2 diabetes” is enough for a helpful answer.
  3. Cover personal details on letters before you photograph them for the AI.
  4. Check your settings. In ChatGPT, you can turn off Improve the model for everyone in Settings > Data controls. See Is ChatGPT safe and free?

A good way to use AI before an appointment

  1. Ask the AI to explain your condition or test in plain words.
  2. Ask it to list questions you might want to ask.
  3. Pick the three that matter most to you and write them down.
  4. Take the list to your appointment, and let your doctor answer them.

Try this

Next time you have an appointment coming up, ask an AI assistant to help you write your questions. You’ll likely walk out understanding more.

Watch out for health scams

Scammers use AI too: fake videos of doctors or celebrities selling “miracle” supplements, and fake health apps. Real doctors don’t sell cures in social media ads. See Medicare scam calls for other health-related scams.

AI for Seniors, Plainly Explained has a full chapter on health appointments and medical letters, with ready-to-use prompts and clear boundaries.