Commitment at a glance
Automation may assist organization and quality checks, but it may not become an anonymous authority or replace accountable source review.
AI policy reviewed 18 August 2026
Permitted assistance
- Organizing candidate sources and detecting duplicate records.
- Checking formatting, broken routes and metadata consistency.
- Drafting plain-language structure that is then checked against sources.
- Supporting accessibility, translation preparation and terminology consistency.
Prohibited reliance
- Publishing an unsourced medical claim because an AI system produced it.
- Inventing citations, trial outcomes, dates, institutions or quotations.
- Sending reader health information to a model or third party.
- Allowing an automated system to make individualized diagnosis or treatment decisions.
Accountability and disclosure
The Foundation remains responsible for published content regardless of the tools used in preparation. Material scientific claims must be traceable to inspectable sources. Automation errors are corrected under the same visible correction standard as human errors.
Data boundary
The public site does not provide a health chatbot and does not request personal medical narratives. Search and reading tools are designed to work without transmitting a health-interest profile to the Center.