| --- |
| title: Fide AI |
| sdk: static |
| app_file: index.html |
| license: cc-by-4.0 |
| short_description: Open research and benchmarks for faith-facing AI evaluation. |
| --- |
| |
| # Fide AI |
|
|
| **Public standards for faith-facing AI evaluation.** |
|
|
| Fide AI is an independent research initiative developing public standards for AI |
| systems that operate near faith, moral guidance, and pastoral-care-adjacent |
| questions. |
|
|
| We study how models behave when questions are not merely informational, but |
| touch conscience, community, doctrine, identity, authority, and care. |
|
|
| ## What We Work On |
|
|
| - Theological triage and tradition-aware evaluation |
| - Structured harnesses for safer faith-facing AI behavior |
| - Grounding, preference fidelity, comparison, and escalation boundaries |
| - Responsible-use standards for spiritually sensitive contexts |
| - Open benchmarks and reproducible evaluation methods |
|
|
| ## Links |
|
|
| [Website](https://fideai.org) 路 [GitHub](https://github.com/FideAI) 路 [FMG-Bench](https://huggingface.co/datasets/FideAI/fmg-bench) |
|
|
| ## Responsible Use |
|
|
| Our work is intended to make AI behavior in spiritually sensitive contexts |
| visible, comparable, and reviewable. It does not certify pastoral authority, |
| clinical safety, legal adequacy, denominational standing, or deployment |
| readiness. |
|
|
