5 Ways AI Can Help Your Health (And 3 Ways It’s Dangerous)

📩 Contact us: [talkingwithdocs@gmail.com]

Do you use AI for your health? Most of us have asked ChatGPT, Gemini or Claude about a symptom, a meal or a lab result. Used the right way, AI is a genuinely useful health tool. Used the wrong way, it can steer you wrong when it matters most.

In this episode of Talking With Docs, Dr. Brad Weening and Dr. Paul Zalzal break down the top 5 ways AI can help you manage your health, the 3 ways you shouldn’t use it because it could be dangerous, and what to watch out for: made-up facts, inconsistent answers, out-of-date information and your privacy.

Why not ask AI for a diagnosis? A 2026 Oxford study in Nature Medicine tested this with 1,298 people. On their own, the chatbots identified the right condition in about 95% of the scenarios. When real people described the same scenarios in their own words, that fell to fewer than 35%, no better than people using an internet search or their own judgment.

New to AI and health? Watch our episode on AI deepfakes: https://youtu.be/rOYudqQ9oq4

You’ll learn:

How AI can plan healthy meals from what’s in your fridge
How to use AI to track macros, even from a photo of a nutrition label
How to get an exercise plan built around your time, fitness, injuries and equipment
How to prepare for a doctor’s appointment with a few focused questions
How AI can explain medical jargon, lab results and imaging reports
How AI can help you build and track healthy habits and routines
Why AI is not good at diagnosing you
Why AI tends to be a people pleaser, and how you can lead it to the answer you want
Why you should never use AI in an emergency
Why you shouldn’t use AI to change a treatment plan your doctor gave you
What AI “hallucinations” are, including made-up references
Why you should challenge AI and ask for its sources
How your privacy changes once your health information leaves the clinic
How to turn off data sharing in your AI settings

The most important rule is about emergencies. Chest pain, shortness of breath, severe abdominal pain, an asthma attack, or stroke symptoms like sudden weakness or vision changes: don’t type it into your phone. Call 911 or get to the hospital. AI might get it wrong or tell you it’s okay, and that’s time you can’t get back.

Dr. Zalzal shares how AI once invented references for a talk he was putting together, including a convincing New England Journal of Medicine article that never existed. So challenge it: ask for its sources, whether the information is recent, and what the argument against it is.

Privacy matters too. In a hospital or doctor’s office, your health information is protected by health privacy laws like PHIPA in Ontario or HIPAA in the US. Once you type it into an app, that company’s privacy policy applies instead, and removing your name doesn’t necessarily make you anonymous. Turn off the setting that lets the model learn from your data.

And give AI context. Instead of “Can I take vitamins with my thyroid medication?”, try “I’m a 58-year-old woman with hypothyroidism. I take this dose of thyroxine and want to start vitamin B12. Is there an interaction?” Ask for plain language and sources. Then bring what you learn to your healthcare team. You’re still in charge of your own health.

How do you use AI for your health? Let us know in the comments.

📄 Study discussed:
Reliability of LLMs as medical assistants for the general public: a randomized preregistered study. Nature Medicine, 2026.
https://doi.org/10.1038/s41591-025-04074-y

🎧 Listen on Your Favourite Platforms:

Spotify:

Apple Podcast:

YouTube Music:
– TWD Podcast

👕 Shop Our MERCH:
https://www.spreadshirt.ca/shop/user/talking+with+docs

❤️ Support Oakville Hospital:

Home Page

📱 Follow Us:

TikTok:

@talkingwithdocs

Instagram:
https://www.instagram.com/talkingwithdocs/?hl=en

🛑 DISCLAIMER: This content is for educational purposes only and should not replace professional medical advice. AI tools can be inaccurate, out of date or inconsistent, and they cannot examine you or know your full health history. Never use AI in place of emergency care, and never start, stop or change a medication or treatment plan based on what an AI tool tells you. Decisions about your diagnosis and treatment should be individualized based on your health history and circumstances. Always follow the guidance provided by your physician, pharmacist, and healthcare team. Full disclaimer at talkingwithdocs.com.

#AI #ArtificialIntelligence #ChatGPT #AIinHealthcare #AIHealth #HealthTech #DigitalHealth #PatientSafety #HealthPrivacy #MedicalAdvice #HealthTips #TalkingWithDocs #DoctorExplains

Share on:

Leave a Reply

Your email address will not be published. Required fields are marked *