Tools do not get the byline

AI use policy

Where assisted technology may help, where it is prohibited, and what remains a human responsibility.

Effective July 31, 2026

Human accountability

A named human editor remains responsible for every published sentence, claim, citation, image, recommendation, disclosure, and correction. An automated system cannot approve health content, invent a source, assign a verdict, clear a legal policy, or press publish.

Permitted assistance

Approved tools may assist with transcription, document organization, formatting, taxonomy suggestions, accessibility checks, code, repetitive data cleanup, brainstorming, and comparison of text against a defined checklist. Output is treated as untrusted until a responsible person verifies it against original sources and policy.

Prohibited uses

We do not publish fabricated quotations, testimonials, people, credentials, testing, citations, images presented as documentary reality, or first-person product experience. We do not use private reader messages, health data, embargoed material, licensed databases, or confidential partner information to train a model without valid authority.

Health and product claims

AI-generated summaries are not evidence. Every material health or performance claim is checked against the primary record, relevant total evidence, and an appropriate human reviewer. Citations are opened and verified; a plausible-looking reference that cannot be confirmed is removed and investigated.

Disclosure and records

Material synthetic media or substantial automation that a reasonable reader would want to know about is disclosed. Editorial records preserve the responsible humans, tool category, source material, review steps, and approval—not raw sensitive prompts or proprietary data.