Day 4 · Week 1 · AI & self-care
AI sounds certain. Here is what I would check.
AI can sound settled even when it is guessing. For my health, I would want less certainty in the delivery and more clarity about what it does not know.
I would use a general health explanation without personal records or identifying details, then use the prompt below to separate general knowledge from claims needing personal context. If it said 'usually nothing to worry about,' I would want to know whether that described a study population or guessed at my situation.
A prompt I'd try:
Copy the prompt, then open ChatGPT and paste it into a new chat.
Review this general explanation: [text without identifying information]. Separate what can be explained generally from what depends on an individual's situation. Name up to two pieces of context a clinician would need before applying it to someone, and explain why. Don't invent that context or assign me a personal risk estimate. Flag claims that need source checking. Keep it brief; confidence in your own answer is not verification.
Copy manually
Press and hold the text below, choose Copy, then paste it in ChatGPT.
I would save missing context and claims needing a source as questions for my doctor. The AI calling its own answer verified would not verify it. “Not urgent” would not establish that waiting was safe. Those open questions would be more useful than a beautifully written answer I could not trust.