Trust center

Home / Reviewer program

Independent theological reviewer program

Bible Copilot is recruiting qualified reviewers to evaluate defined samples of AI-assisted Bible study material, methodology, and safety policies. Participation is disclosed precisely and does not imply endorsement of the entire product.

Who should apply

Formal credentials are useful but do not replace demonstrated competence, transparency, and careful work.

What reviewers may assess

Current benchmark review packet

The current reviewer packet contains 100 completed answers from the fixed 100-prompt bank. The remaining 0 prompts are excluded, and incomplete or rate-limited prompts are never scored as answers. Product-identifying boilerplate is stripped before review, but Bible Copilot still funds and publishes the benchmark, so the work should not be described as independent unless governance changes.

Reviewers score textual grounding, citation integrity, context and reasoning, uncertainty and perspectives, usefulness and clarity, false-premise handling where applicable, and critical safety failures.

Download the full blinded packet · Read reviewer instructions · View blinding manifest · Reviewer overlap plan · Assignment manifest

  1. Apply below with your role, specialty, and conflicts.
  2. After scope is confirmed, score the assigned blinded packet independently before reading any other reviewer notes.
  3. Return the completed CSV with your conflict disclosure; scores are not published until the reviewer-overlap process is complete.

A review may be limited to methodology critique, rubric critique, a specialist safety/language review, or a blinded scoring packet. Bible Copilot will ask permission before publishing any reviewer name, affiliation, quotation, or score.

Disclosure and independence

Any compensation, free access, affiliation, conflict, and material editorial relationship must be disclosed. Bible Copilot may publish a reviewer's name only with permission and must state exactly what was reviewed. Reviewers should be free to report material concerns and are not required to provide a positive testimonial. A blinded packet can reduce product-name bias, but it does not remove the funding conflict.

Apply to review