Why we have a code
Infection preventionists and clinicians act on what our tools tell them, and patients and residents live with the results. These eight commitments are what we hold ourselves to every time Prevexia answers a question, reviews an inspection or drafts a report, and we review new AI features against them before they reach you.
1. Do no harm first
When Prevexia is unsure, it says so and points you to a person or an authoritative source.
2. Privacy is absolute
Never enter patient information. Our tools screen for it, and we collect only what we need to run the service.
3. Truth with sources
Prevexia shows the guidance its answers come from, and tells you when it cannot find a source.
4. You decide
Prevexia supports clinical judgment. It never replaces it, and it never acts on a patient's care by itself.
5. Fair to every setting
A small rural nursing home deserves answers as good as a large hospital's. We look for gaps and close them.
6. Always disclosed
You always know when you are talking with AI, and you can see why it answered as it did.
7. Accountable
We review problems, fix them, and tell you about the ones that affect you.
8. Stay in scope
Prevexia does not diagnose patients or give legal advice.
What this code is built on
- The four principles of biomedical ethics: beneficence, non-maleficence, autonomy and justice
- ANA Code of Ethics for Nurses with Interpretive Statements
- AMA Code of Medical Ethics
- WHO, Ethics and governance of artificial intelligence for health (2021)
- NIST, AI Risk Management Framework 1.0 (2023)
- ONC, HTI-1 final rule (2024): decision support transparency
- FDA, Clinical Decision Support Software guidance (2022)
- HIPAA Privacy and Security Rules
Tell us when we fall short
If an answer or a feature seems to break one of these commitments, email hello@prevexium.com and tell us what you saw. Please do not include patient information. A person reads every message. See also our Disclaimers & Notices and Security pages.
Related policies