Dr Philip Crilly’s ASK SAFE Framework for AI Health Advice
A Practical Framework For Safer AI Health Advice
Many people ask AI health questions without providing enough context.
ASK SAFE (version 1.1) is a structured framework designed to help people have more complete and useful conversations with AI when seeking health information.
A = Age & Background
Include information about the person affected, such as:
- Age
- Medical conditions
- Medications
- Relevant background information
S = Symptoms
Describe:
- Main symptom
- Duration
- Severity
- Whether it is improving or worsening
- Any accompanying symptoms
- Experienced anything similar before
K = Knowledge Sources
Ask AI to base its response on recognised, evidence-based sources where appropriate (for example, NHS or NICE guidance in the UK).
S = Safety Questions
Ask:
What red flags should I watch out for?
A = Ask Follow-Up Questions
Ask:
What follow-up questions would you ask before giving advice?
F = Find The Right Care
Ask:
Based on the information available, what level of care may be appropriate?
Examples:
- Self-care
- Pharmacist
- GP
- NHS 111
- Emergency care
E = Escalation Plan
Ask:
What should I do if symptoms change, worsen or fail to improve?
Important Reminder
ASK SAFE is designed to support understanding and decision-making.
AI can make mistakes and may not have all the information needed to assess your situation. If you are seriously unwell or concerned, seek advice from an appropriate healthcare professional or emergency service.
Dr Philip Crilly 