Choose your role and situation to find questions for a real conversation.
Ask
Choose your role, then choose a situation. You will see prepared questions you can copy and take to a visit, a team meeting, or a conversation at home.
A health system is not a market first. It is people who will live with what you ship. Translation is fitting the tool to that place.
For I am building or bringing a tool into care
A way to begin
What would we need to know before we let this rank, write, or watch a person in care?
Who was this tested on, and do they look like the people who will use it?
What must remain a human judgement even if our model is faster?
If this fails, is it a nuisance, a delay, or an irreversible harm, and who is named?
What can a patient or a clinician refuse, without the rest of care getting worse?
Where does the data go, including audio and notes, and who is the controller?
Triage algorithm bias
Triage algorithm bias occurs when an automated urgency score gives some groups lower priority than their medical needs justify.
AI documentation "ambient scribes"
Ambient-scribe risk arises when a tool listening to a clinical conversation creates an inaccurate record or captures and uses speech in ways the patient did not understand.
Diagnostic errors
Diagnostic error happens when AI suggests the wrong explanation for a person's symptoms or fails to identify an illness that is present.
These resources offer more detail. Each title opens the original source in a new tab.
Health-specific guidance
World Health Organization
WHO guidance on the ethics and governance of large multimodal models in health.
Use this when considering generative AI in a health or care setting.
How to manage, not only name
U.S. National Institute of Standards and Technology
A voluntary framework for organizing work to identify, assess, and manage AI risks.
Use this when your organization needs a process for managing AI risk.