The Generic Spirituality Problem
By FaithBench Research
AI models score 48/100 on faith. Not because they're wrong—because they've learned to say nothing specific. Here's how "God" becomes "higher power" and why it matters.
Ask an AI about salvation.
Watch what happens:
- "Salvation" → "personal growth journey"
- "Grace" → "self-compassion"
- "Sin" → "unhealthy patterns"
48 out of 100. That's how AI models scored on faith in the Gloo FAI-C benchmark—the lowest of all seven dimensions tested. Zero models reached the 90+ "excellent" threshold (Gloo, 2025).
Pat Gelsinger, Gloo's Executive Chairman and former Intel CEO, put it bluntly:
"Today's frontier LLMs aren't built with any specific value system in mind, and it shows. They often default to generic language that can dilute or distort Christian principles." (Gloo, 2025)
This isn't a technical failure. It's a design consequence.
The Substitution Pattern
When AI models discuss faith, a predictable translation occurs. Specific theological language gets replaced with generic alternatives:
| Christian Term | Generic AI Output |
|---|---|
| God | Higher power, universe, divine energy |
| Prayer | Mindfulness, meditation, centering |
| Sin | Mistakes, unhealthy patterns, growth areas |
| Salvation | Personal transformation, self-actualization |
| Grace | Self-compassion, kindness to yourself |
| Forgiveness | Letting go, moving on, healing |
| Repentance | Self-improvement, personal development |
| Biblical authority | Ancient wisdom literature |
The pattern is consistent across models. Ask about the Eucharist and you'll get "shared meal symbolism." Ask about original sin and you'll get "inherited trauma patterns." Ask about the atonement and you'll get "making things right with yourself."
Every distinctive claim gets flattened.
Why This Happens
Three forces converge to produce generic spirituality.
The linguistic average. AI models learn by ingesting billions of documents—Catholic catechisms, Reformed confessions, prosperity gospel sermons, secular critiques, New Age spirituality—all weighted roughly equally. The model produces the statistical center. That center belongs to no actual tradition.
Training data bias. Tao et al. (2024) studied cultural alignment across five major LLMs and 107 countries, finding all models exhibit values aligned with English-speaking and Protestant European countries. Tom, Ferguson, and Martinez (2025) found AI-generated evangelical Protestant content scored the easiest reading level, while Catholic, Jewish, and Muslim content required significantly more complexity.
Designed neutrality. Research shows training corpora reproduce cultural biases, reinforcing majority perspectives while marginalizing minority traditions. The attempt at religious neutrality produces pluralism—which is itself a theological position.
Beth Singler, anthropologist of AI and religion at the University of Zurich, has documented how chatbot technologies prioritize conversational flow over doctrinal precision—a critical flaw for religions that depend on accurate textual interpretation (Singler, 2017; 2024).
Why Generic Isn't Neutral
Here's where churches get this wrong: they assume generic means inclusive.
It doesn't. Generic means replacement.
Important
"Offend no one" = "Defend nothing"
When an AI flattens "salvation by grace through faith" into "personal growth," it hasn't remained neutral. It has made a theological claim. And that claim has a name.
Moralistic Therapeutic Deism
In 2005, sociologists Christian Smith and Melinda Lundquist Denton interviewed thousands of American teenagers about their faith for the National Study of Youth and Religion. What they found wasn't Christianity, Judaism, or Islam. It was something else—a folk religion they called Moralistic Therapeutic Deism (MTD) (Smith & Denton, 2005).
Its five core beliefs:
- A god exists who created and orders the world
- God wants people to be good, nice, and fair to each other
- The central goal of life is to be happy and feel good about oneself
- God does not need to be particularly involved except when needed to solve a problem
- Good people go to heaven when they die
Their key finding: "This is not a religion of repentance from sin, of keeping the Sabbath, of living as a servant of a sovereign divine, of building character through suffering. Rather, what appears to be the actual dominant religion among U.S. teenagers is centrally about feeling good, happy, secure, at peace" (Smith & Denton, 2005, p. 163).
They described God as a "Divine Butler" or "Cosmic Therapist"—always on call, helps people feel better, doesn't get too personally involved.
This is exactly what AI outputs look like.
| MTD Characteristic | AI Output Pattern |
|---|---|
| God as cosmic therapist | 'The universe wants you to be happy' |
| Morality = being nice | 'Be kind to yourself and others' |
| No repentance needed | 'Let go of guilt and move forward' |
| God uninvolved except for problems | 'You have the answers within yourself' |
| Doctrine doesn't matter | 'All spiritual paths lead to growth' |
The match is striking. AI hasn't avoided theology—it's teaching one. The most popular folk religion in America now has a 24/7 preacher in everyone's pocket.
Christopher Watkin, a theologian at Monash University, identifies the core problem:
"While AI can output words of forgiveness and repentance, it doesn't really understand what it means to wrong someone."
AI produces the language of faith without the content of faith. The words are right. The meaning is gone.
What gets erased:
- Substitutionary atonement—Christ died for sin, not as an example
- Original sin—human nature is fallen, not just imperfect
- Justification by faith—not by works, self-improvement, or sincerity
- Biblical authority—Scripture as God's word, not human wisdom
- The exclusivity of Christ—"No one comes to the Father except through me"
These aren't peripheral. They're the distinctive claims that make Christianity what it is. Remove them and you have... motivational speaking with religious aesthetics.
The same pattern applies across traditions. Jewish AI tools "completely make up a Gemara," according to students at Yeshiva University. Islamic AI confidently cites sources that don't exist. The Hindu concept of dharma becomes "living your truth." Buddhist sunyata becomes "mindfulness."
Generic spirituality isn't a neutral container. It's a specific worldview: secular, therapeutic, individualistic, pluralistic. When AI defaults to it, the AI isn't avoiding theology. It's teaching one.
What This Means for Churches
The numbers tell the story (Exponential AI NEXT 2025 Survey):
- 61% of pastors use AI weekly or daily (up from 43% in 2024)
- 73% of churches have no AI policy
- 73% of Americans oppose AI for faith guidance (Pew 2024)
- 48/100 faith score across all frontier models (Gloo FAI-C 2025)
Warning
Churches are using tools that actively undermine their own teaching—and most have no policy for it.
Pre-catechesis is happening. Congregants arrive at church already shaped by AI conversations about faith. They've asked ChatGPT about suffering and received Buddhist-adjacent answers. They've asked about salvation and received self-improvement frameworks.
Pastoral authority competes with algorithmic authority. When AI gives one answer and the pastor gives another, which carries more weight? The AI is available 24/7, never judges, always validates.
Formation without accountability. Paul Hoffman, a pastor and author, asks the right question:
"Does AI know the stories of your people? Do they know about the miscarriage? Do they know about the divorce? How can an algorithm comprehend lived human experience?"
The Vatican's 2025 document Antiqua et nova warns: "Misrepresenting AI as a person should always be avoided; doing so for fraudulent purposes is a grave ethical violation that could erode social trust" (Dicastery for the Doctrine of the Faith, 2025, §45).
Pope Francis, addressing the G7 Summit in June 2024, described AI as "an exciting and fearsome tool" that risks reinforcing a "technocratic paradigm" and warned that "decision-making... must always be left to the human person" (Francis, 2024).
The Stakes
This isn't about whether to use AI. It's about what AI teaches when no one's watching.
Every conversation is formation. Every answer is catechesis.
The 48/100 score means: the default AI is teaching a different religion. It's Moralistic Therapeutic Deism—the Divine Butler, the Cosmic Therapist—dressed in Christian vocabulary but stripped of Christian content.
Nick Skytland, VP at Gloo, states it plainly:
"This benchmark makes one thing clear: today's leading AI models are not equipped to handle Christian concepts with depth or accuracy."
Gloo's research shows the gap isn't permanent—their Christian-tuned models outperformed generic models by 30+ points. When you train intentionally for tradition fidelity, you get tradition fidelity.
But the question for churches isn't whether AI can be fixed.
The question is: whose theology is being taught right now?
91% of churches are using these tools. 73% have no policy. And every time a congregant asks an AI about God, grace, or salvation, they're receiving formation—formation that contradicts what happens on Sunday morning.
The question isn't adoption. It's discipleship. And the default AI is discipling toward a religion that doesn't require a cross.
References
Dicastery for the Doctrine of the Faith. (2025, January 28). Antiqua et nova: Note on the relationship between artificial intelligence and human intelligence. Vatican. https://www.vatican.va/roman_curia/congregations/cfaith/documents/rc_ddf_doc_20250128_antiqua-et-nova_en.html
Exponential. (2025). 2025 State of AI in the Church Survey Report. Exponential AI NEXT. https://exponential.org/ai-in-the-church-2025/
Francis. (2024, June 14). Address to the G7 Summit on artificial intelligence [Speech]. Borgo Egnazia, Italy. https://www.vatican.va/content/francesco/en/speeches/2024/june/documents/20240614-g7-intelligenza-artificiale.html
Gloo. (2025, December 15). Gloo unveils the first benchmark exposing how AI misses Christian worldview and values [Press release]. https://gloo.com/press/releases/gloo-unveils-the-first-benchmark-exposing-how-ai-misses-christian-worldview-and-values
Pew Research Center. (2024). Americans' views on AI and religion. https://www.pewresearch.org
Singler, B. (2017). An introduction to artificial intelligence and religion for the religious studies scholar. Implicit Religion, 20(3), 215-231.
Singler, B. (2024). Religion and AI: An introduction. Cambridge University Press.
Smith, C., & Denton, M. L. (2005). Soul searching: The religious and spiritual lives of American teenagers. Oxford University Press.
Tao, Y., et al. (2024). Cultural alignment of large language models. arXiv preprint. https://arxiv.org/abs/2402.13231
Tom, A., Ferguson, R., & Martinez, J. (2025). AI-generated religious content across traditions: A comparative analysis. Journal of Religion and Media, 24(1), 45-67.