About
A research lab for faith-facing AI.
Fide AI exists to study how AI systems reason, retrieve, and respond when people ask questions of theology, moral formation, religious education, and pastoral-adjacent care.
AI should raise human dignity, not erode it. In faith-facing contexts, that requires more than good intentions; it requires evidence.
I started Fide AI because AI systems are already entering spaces where people ask sacred, moral, and pastoral-adjacent questions, but the public evidence base is still thin. Too much of the conversation relies on intuition, anecdotes, or generic AI safety language that was not designed for theology, formation, religious education, or pastoral care.
The work grows out of both technical and lived commitments: years building and evaluating AI systems, and a conviction from church life that questions of doctrine, dignity, agency, love, and care deserve careful handling. Fide AI exists to make those questions measurable, inspectable, and accountable before high-trust institutions rely on AI systems.
Why now
AI systems are moving into high-trust domains before the field has a serious research base for religious and moral use. The question is no longer only whether AI can answer religious questions. It is whether AI systems shape people's trust, agency, humility, relationships, and moral imagination in ways institutions can inspect.
What makes faith-facing AI hard
- Theological claims can be high-stakes while remaining tradition-specific.
- Moral guidance must navigate genuine diversity without collapsing it.
- Pastoral contexts require safety awareness that most AI evaluation ignores.
- Faith communities have deep history with authority that AI can accidentally mimic.
- AI systems can simulate intimacy or spiritual authority without the accountability of embodied community.
- Grounding matters: invented scripture or history can cause real harm.
Institutional role
Fide AI is building research infrastructure for institutions that need to understand AI behavior before adoption. Product companies, funders, reviewers, and faith institutions can participate under published rules, claims limits, and independence controls.
Broader than benchmarks
FMG-Bench is the first public benchmark release, not the whole mission. The broader agenda includes theological reasoning, retrieval and grounding, comparative tradition representation, human dignity and formation, pastoral-adjacent safety, institutional readiness, and governance for high-trust use.