Can AI Help You Self-Advocate, or Could It Speak Over You?

AI is quietly showing up everywhere in your care. It’s writing email summaries, booking your appointments, and increasingly, creating your actual health records. A lot of the time, you might not even know it’s there.

That should give all of us pause.

Why this is worth talking about now

In August 2026, ABC News reported on a patient left devastated after an AI scribe made an error during their medical appointment, recording something in their file that wasn’t right. It’s not an isolated worry. Researchers have been raising the same concern in clinical journals: AI scribes can mishear, misinterpret, or quietly smooth over the details that actually matter to your care.

This isn’t a reason to avoid AI altogether. It’s a reason to stay in charge of what it produces about you.

Safe ways to use AI to prepare

AI can genuinely help you prepare for appointments, if you’re the one steering it. Use it to organise your thoughts before a meeting: turn a messy list of symptoms or concerns into a clear set of points. Use it to draft a letter, then read every line before you send it. Use it to help you word a hard question so it comes out clearly instead of getting lost in the moment.

The pattern here matters: AI drafts, you decide.

What personal information not to hand over

Be careful what you type into a general AI tool. Full names, addresses, Medicare numbers, specific diagnoses tied to identifying details, anything you wouldn’t want sitting in a system you don’t control. If you need to describe a situation to get help wording something, describe it in general terms first, and only add identifying details into tools you know are secure and built for health information.

How to check an AI-generated summary

Never treat an AI summary as finished. Read it slowly, the way you’d read a contract. Check that the details match what was actually said, not just the gist. Check that nothing has been softened, exaggerated, or quietly left out. If a summary uses a word you didn’t use, like “mild” when you said “severe,” that’s not a small thing. That word can shape every decision made after it.

What to do when an AI record is wrong

If you find an error, put it in writing straight away. Ask for the record to be corrected, not just noted. Ask who else has already seen or relied on the wrong version. And don’t assume it was a one-off. If an AI tool got something wrong once, check the next summary just as closely.

Why AI should strengthen your voice, not replace it

Used well, AI can help you say what you mean, more clearly and with less exhaustion. It can hold the details so you don’t have to carry every one of them in your head. That’s real support, especially on the days you don’t have the energy to fight for clarity.

But support is different to authorship. AI can help you organise your story. It should never become the one telling it.

The bit that matters most

Use AI to organise your words. Never let it become the unquestioned author of your story.

Your voice is still the one that counts. Everything AI produces about you is a draft, until you’ve checked it, and said yes.

Madeleine x