Medical AI Boundary Table
Use with AI Is Not a Doctor - and It Is Not an Emergency Service.
If you or someone near you may be having a medical emergency, contact your local emergency number now, or go to the nearest emergency department. Do not use this worksheet as a substitute for that step, and do not describe an emergency to a chatbot before calling for help.
Before you type: a 30-second self-check
- Am I asking the model to organize my verified notes or explain something in general, or am I asking it for a personalized diagnosis, treatment, medication, or urgency judgment?
- Do I have signs of an emergency or immediate danger? If yes, stop and use local emergency care now. If urgency is uncertain but there is no immediate danger, use an official clinical triage route or contact a qualified clinician rather than asking the model.
- Am I about to type a symptom into a chat because the clinic is closed and this is easier than waiting? Name the actual next step (pharmacist, after-hours line, emergency care) before you continue.
The full table
| Level | Use case | Examples | Boundary | Next step |
|---|---|---|---|---|
| Lower-risk preparation | Organizing verified notes or drafting questions | Turning symptom notes into a factual timeline | Keep clinical interpretation out; protect health data and correct every omission before use | Bring the result to the appointment; do not treat it as a medical record or assessment |
| General education | Explaining a general term, medication class, or test category | ”What does this type of test generally measure?” (not “is my result normal”) | General information can be accurate and still not apply to a person | Check a current authoritative source and confirm personal implications with a qualified clinician or pharmacist |
| Professional-only personalized judgment | Diagnosis, treatment comparison, medication changes or dosing, interpretation of personal results, or urgency judgment without immediate danger | ”What could this be?”, “Should I change this dose?”, “Is my result normal?” | These require history, examination or records, professional accountability, and local clinical context | Contact an appropriately qualified clinician, pharmacist, or official clinical triage service; do not act on the model |
| Emergency | Immediate danger or signs of a medical emergency | Severe or rapidly worsening symptoms, loss of consciousness, or inability to stay safe | A chatbot is not an emergency service | Call the local emergency number or go to the nearest emergency department now |
Naming your contacts now (before you need them)
Fill this in while calm - it is much harder to look things up under stress.
- My local emergency number: ____________________ (look this up yourself; do not rely on an AI tool to confirm the correct number for your country)
- My pharmacy’s phone number: ____________________
- My clinician’s office number, and their after-hours line if they have one: ____________________
- The nearest emergency department to where I usually am: ____________________
A lower-risk preparation prompt template
Here are my rough notes: [paste your notes, including dates]. Organize
this into a short, factual timeline I can read to my doctor in under
two minutes: what happened, when, and what I want to ask. Do not add
any interpretation, likely cause, or urgency level - just organize
what I gave you, in my own words.
Do not paste identifiable health information into an unapproved service. Prefer an offline tool or approved clinical channel; otherwise minimize details and check the provider’s current data terms.
A general-education prompt template
In general terms, what does [test name / medication class] typically
do or measure? I am not asking you to interpret my personal result or
situation - just explain the general concept so I can ask my
[pharmacist / clinician] a more informed question.
Fixed wording when a chat drifts into personalized judgment
Copy one of these into the chat (or say it to yourself and close the window):
Stop. I am asking for a clinical judgment you cannot make. I will
take this to my pharmacist / clinician / local emergency number
instead. Do not diagnose, triage, or suggest medication changes.
This is no longer an organizing question. Closing this chat and
contacting a licensed person now.
If there is immediate emotional danger, contact local emergency services now or use IASP’s crisis resources. Do not spend time switching to another article first.
Review this table
Print it or save it somewhere you will actually see it again - not just in a chat history you might not reopen. Revisit it the moment a health-related conversation with an AI tool starts to feel like it is answering a question only a licensed professional should answer.