AI Has a God Problem Stretching From the Vatican to the Baptist Pulpit
By Gammatek ISPL Last updated: October 2026 | 14 min read
Author note: This piece draws on publicly reported Vatican documents, named Church officials' public statements, and independent reporting on AI governance. Where a claim is attributed to a specific source, that source is named in the text.
Why This Matters
When people picture who's shaping the rules for artificial intelligence, they usually picture regulators in Brussels, engineers in San Francisco, or lawmakers in Washington. Few picture a Franciscan friar with an engineering doctorate sitting on a United Nations AI advisory board, or a newly formed Vatican commission with its own encyclical on the subject. But that's exactly what's happening — and it matters to anyone building, deploying, or simply living alongside AI systems, because religious institutions reach billions of people's everyday moral reasoning in a way that a corporate ethics board or a government white paper never will. If you're trying to understand where public trust in AI will actually come from — not just where the technical safeguards sit, but where the permission structures for using this technology get built — you need to understand what's happening in pulpits and Vatican commissions, not just in boardrooms.
The Vatican Isn't Just Commenting on AI — It's Formally Governing It
In January 2025, the Vatican's Dicastery for the Doctrine of the Faith, jointly with the Dicastery for Culture and Education, issued a formal doctrinal note titled Antiqua et nova ("Ancient and New"), examining the relationship between artificial intelligence and human intelligence and laying out the Church's position on the ethical challenges AI raises. This wasn't a passing comment from a single bishop — a doctrinal note is an official instrument of the Church's teaching authority.
That same year, Time magazine placed Pope Leo XIV on its list of the world's most influential people in artificial intelligence — a striking inclusion alongside tech CEOs and AI researchers, and a signal that mainstream coverage now treats the papacy as a genuine actor in the AI governance conversation, not a bystander commenting from the sidelines.
Then, in May 2026, Pope Leo XIV approved the creation of a dedicated Vatican commission on artificial intelligence and released his first encyclical — Magnifica humanitas — focused specifically on AI's relationship to human dignity. An encyclical is among the highest forms of papal teaching; devoting the first one of a papacy to AI is a deliberate, significant signal about where Church leadership believes its moral authority needs to be exercised next.
Meet the Friar Advising the Vatican on AI
Behind these documents is a specific person worth knowing: Friar Paolo Benanti, a Franciscan priest with an engineering background and a doctorate in moral theology, who teaches at the Pontifical Gregorian University. He's described in reporting as the Vatican's go-to figure on technology, with direct access to papal leadership on the subject. His influence extends well beyond the Church — he sits on the United Nations' Advisory Body on Artificial Intelligence, heads an Italian government commission on disinformation and journalism, and consults for the Vatican's Pontifical Academy for Life.
Benanti's framing of the AI problem is notably pragmatic rather than restrictive. He has described the core challenge as "a problem of governance" — not a question of whether AI should exist, but of finding the right level of its use within a given social context. He sees real benefit in the technology, citing its potential to lower medical costs and help doctors reach more patients. But he also raises two concerns that get far less attention in mainstream tech coverage:
Non-inclusive training data produces non-inclusive outcomes — a technical point made in explicitly moral language, tying data governance directly to questions of justice.
The labor behind AI systems matters morally, not just legally. Benanti has specifically criticized the use of underpaid workers — often in countries that are former colonies — to perform the data-labeling and content-moderation work that trains AI systems, saying he does not want this era of AI development remembered as one that extracted cognitive labor from the Global South.
He has also argued that AI regulation shouldn't aim to limit development outright, but should keep the technology compatible with what he calls the "fragile system that is democracy" — a framing that treats AI governance as fundamentally about protecting democratic institutions, not just protecting individual users.
Why This Is Different From Corporate AI Ethics Statements
Most AI ethics content people encounter comes from the companies building the technology — safety pages, responsible-AI frameworks, model cards. These are useful, but they share a structural limitation: the organization writing the ethics statement also has a direct commercial interest in how permissively that statement gets interpreted.
Religious institutions occupy a genuinely different position. The Vatican has no product roadmap, no shareholder pressure, no competitive incentive to ship faster than a rival. That doesn't make its judgment automatically correct — a 2,000-year-old institution reasoning about large language models will inevitably bring its own blind spots and priors. But it does mean the incentive structure producing Antiqua et nova is fundamentally different from the one producing any single AI company's safety framework, and that difference is part of why it carries distinct weight in the public conversation, independent of whether you share the Church's theology.
It's Not Just the Vatican: Protestant and Evangelical Leaders Are Having Their Own Version of This Debate
The same underlying tension — how much authority, trust, and intimacy should be extended to an AI system — is playing out in American Protestant and evangelical communities too, though with less centralized institutional structure than the Vatican's.
Baptist and evangelical leaders have increasingly had to address a question that didn't exist in any serious way five years ago: what happens when congregants turn to AI chatbots for spiritual guidance, prayer, or even a version of pastoral counseling? Reporting on the use of AI for religious purposes has documented a growing, genuine pattern of people using general-purpose chatbots to "communicate with God" — seeking comfort, moral guidance, or a stand-in for pastoral conversation they either can't access or are uncomfortable seeking from a human.
This raises a distinctly different set of concerns than the Vatican's governance-focused framing. Where Benanti's Vatican-level work focuses on labor, inclusivity, and democratic compatibility, evangelical and Baptist leaders are often confronting a more immediate pastoral question: can an AI system, however articulate, actually fulfill functions historically understood as requiring human presence, accountability, and relationship — confession, communal worship, pastoral counsel in moments of crisis? Without a centralized teaching authority equivalent to the Vatican's, these conversations are happening more unevenly, congregation by congregation and denomination by denomination, but the underlying anxiety is strikingly similar: that the technology is moving faster than the institutions responsible for providing moral guidance about it.
A Comparison: Two Religious Traditions, Two Different AI Anxieties
Catholic Church (Vatican) | Baptist / Evangelical traditions | |
Governance structure | Centralized — doctrinal notes, papal encyclicals, a formal commission | Decentralized — individual congregations and denominational bodies respond independently |
Primary framing | Governance, labor justice, data inclusivity, democratic stability | Pastoral authenticity — can AI fulfill roles requiring human relationship and accountability |
Key public figures | Pope Leo XIV, Friar Paolo Benanti | No single unifying figure; distributed across denominational leadership |
Formal output | Antiqua et nova (2025), Magnifica humanitasencyclical (2026), dedicated AI commission | Largely sermons, denominational statements, and pastoral guidance — less formalized |
Core question | How should society govern AI's development and deployment? | Can — and should — AI stand in for human spiritual relationship? |
What This Means for Anyone Building or Deploying AI Systems
This isn't purely an academic or theological curiosity — it has direct implications for how AI products get designed, trusted, and adopted, especially for products that touch emotionally or morally sensitive contexts.
1. "Does this system know its limits?" is becoming a real design question, not just a philosophical one. Benanti's "governance over prohibition" framing — finding the right level of AI use for a given context — is directly applicable to product design: a chatbot deployed for general customer support needs very different guardrails than one that ends up, intentionally or not, being used for emotional or spiritual support. Systems that don't distinguish between these contexts inherit the Baptist pastoral concern whether their designers intended it or not.
2. Training data inclusivity isn't just a technical metric — it's being framed publicly as a moral one. Benanti's point about non-inclusive data producing non-inclusive outcomes is a claim that's now being made from pulpits and Vatican documents, not just AI research papers. Organizations that treat data inclusivity purely as a technical checkbox are increasingly out of step with how a meaningful share of the public is being taught to evaluate AI ethically.
3. Labor practices behind AI systems are now part of the public ethical conversation, explicitly. Benanti's criticism of underpaid Global South data-labeling work is a concrete, citable example of a moral authority connecting AI's supply chain to its ethics — a framing that puts real pressure on any organization's AI vendor and labor practices, not just its model outputs.
4. Trust in AI is not going to be won by technical safety claims alone. If a meaningful part of the public's moral reasoning about AI is being shaped by religious institutions rather than corporate safety pages, then companies and software vendors who only ever speak the language of benchmarks and red-teaming are missing an entire register of the conversation their actual users are having.
An Implementation Consideration: What "Governance" Actually Looks Like in Practice
Benanti's framing — that the core issue is governance, not prohibition — is easy to agree with in the abstract and much harder to operationalize. In practice, "the right level of AI use within a social context" requires something concrete: documented policies about where a given AI system is and isn't appropriate to deploy, clear escalation paths to human oversight when a system is being used outside its intended context, and genuine accountability when those boundaries are crossed — not just a disclaimer buried in terms of service.
This is, not coincidentally, close to the same discipline that industrial compliance frameworks require in entirely different domains: a system that's powerful enough to cause harm needs documented boundaries, auditable oversight, and a clear chain of accountability — whether that system is a chatbot being used for pastoral counseling or a piece of industrial equipment on a factory floor. The specific content of the rules differs enormously; the underlying discipline of "define the boundary, document the oversight, make accountability real" does not.
Where This Goes Next
The Vatican's AI commission and its first dedicated encyclical suggest the Church intends to keep playing an active, ongoing role in this conversation rather than treating Antiqua et nova as a one-time statement. Expect more formal doctrinal engagement as AI capabilities continue to expand, particularly around questions Benanti has already signaled as priorities — labor practices in AI's supply chain and AI's relationship to democratic institutions.
On the Protestant and evangelical side, expect the current unevenness to persist for longer, simply because there's no equivalent centralized body to issue a unifying statement — but also expect individual denominations and influential pastors to increasingly publish their own guidance as AI chatbot use for spiritual purposes continues to grow, since the pastoral pressure driving that conversation isn't going away.
For anyone building AI products, the practical takeaway is this: the public's moral vocabulary for evaluating AI is being shaped well beyond Silicon Valley's own framing, by institutions with centuries of experience reasoning about authority, trust, and the limits of what any system — human or artificial — should be permitted to do. Ignoring that conversation doesn't make it less relevant to how your product gets trusted or distrusted.
Governance Discipline Applies Beyond Theology
Whether the system in question is a chatbot being asked to stand in for pastoral counsel or an industrial process being monitored for safety compliance, the same underlying discipline applies: define what the system is and isn't permitted to do, document who's accountable when it's used outside that boundary, and make sure oversight is real rather than theoretical. That's the practical, unglamorous work behind every governance framework that actually holds up under scrutiny — religious or industrial.
[See how Gammatek's compliance platform builds auditable oversight and accountability into industrial AI and automation deployments → https://www.gammateksolutions.com/post/it-s-all-fun-and-games-until-you-give-ai-your-credit-card https://www.gammateksolutions.com/post/top-mathematicians-are-outraged-by-openai-s-methods
Comments