The short version
Most articles on Edifi are drafted with the help of AI (specifically Claude, made by Anthropic). A human editor reviews every article before it’s published, decides whether it ships, and is responsible for what we put under your name and ours. AI helps us cover ground faster. It doesn’t get the last word on what we publish.
What AI does in our pipeline
The model is used to:
- Draft the first version of long-form articles from a structured editorial brief.
- Pick relevant topic and tag labels from our closed editorial vocabulary.
- Suggest internal links between related articles on the site.
- Generate response posts when we’re commenting on a specific piece of social-media content.
The brief includes the topic, the audience, the source material (when responding to a piece of social content, the actual post text), and our editorial constraints: voice, theological commitments, the kinds of citations we expect, what we will and won’t claim.
What humans do
- Editorial review.A human reviewer reads every article before it goes from draft to published. Articles that don’t meet our quality bar are revised or rejected.
- Theological accuracy.Reformed Christian theology is the editorial foundation. The reviewer verifies that Scripture citations are correct, that attributions to specific theologians are accurate, and that the article doesn’t flatten doctrine for clarity.
- Mental-health accuracy.Where an article touches on mental health, we ensure it doesn’t make clinical claims, doesn’t replace professional care, and acknowledges the limits of what a written piece can do.
- Source verification.If the article cites a statistic, study, book, or person, the reviewer verifies the citation. We don’t publish numbers we can’t source.
- Editorial voice. The reviewer ensures the article reads as Edifi: warm, clinically credible, pastorally honest, never preachy or corporate.
What we won’t do
- We will not publish AI-drafted content without human review.
- We will not put a real human’s name on an article as the primary author when the article is AI-drafted. The byline reflects the human who reviewed and stands behind the piece; AI assistance is disclosed on every article.
- We will not use AI to invent statistics, studies, dates, or biographical details. If we cite a number, that number came from a real source.
- We will not use AI to generate fake quotes, fake testimonials, or fake counselor profiles.
- We will not use AI to give clinical mental-health advice. AI-drafted articles are general educational content; they are not therapy, diagnosis, or treatment.
How we disclose AI use
Every AI-assisted article ends with an editorial note that names the AI model used (Claude, by Anthropic), confirms the article was reviewed by a human editor, and links back to this policy. The note appears above the standard care-disclaimer at the foot of the article.
This is the same standard the Federal Trade Commission and Google’s search-quality team recommend for AI-assisted content: disclose, don’t imply human-only authorship, take editorial responsibility.
Why we use AI at all
Christian counseling sits at the intersection of historic Reformed theology and modern mental-health care, and there’s a real shortage of writing that takes both seriously at once. AI helps us cover more of that ground than a small editorial team could on its own. The bar we hold ourselves to: every article should be as theologically careful and clinically careful as if a human wrote it from scratch. If AI assistance makes that bar harder to clear on a given piece, we don’t publish it.
Questions, feedback, corrections
If you spot an error, suspect a fabricated citation, or want to push back on something an AI-assisted article says, please email us:
Edifi, LLC
Email: hello@tryedifi.com
We read every message and reply within a few business days.
Last Updated: May 22, 2026