Religion and AI Ethics: Who Decides What Machines Are Allowed to Do?

When a major religious leader publishes 42,000 words on artificial intelligence, secular tech leaders pay attention. Whether they admit it or not, the public response inside large AI labs to the Vatican's positions on AI has been more careful than the response to most other external commentary. The reason is not piety. It is that religion remains, for billions of users, the working vocabulary for moral choice — and AI is a technology whose moral questions cannot be answered by engineering alone.

Why the religion question keeps appearing

AI governance was supposed to be a secular project. Engineers, lawyers, ethicists trained in academic philosophy, regulators. The shape of the field was set by people for whom religious vocabulary was at best private. But the questions that AI raises — what counts as a person, when a decision deserves human dignity, who is responsible when a system causes harm, what to do with end-of-life choices an algorithm makes faster than a family can — keep landing on territory that religious traditions have been working through for two thousand years.

You can refuse to consult religious frameworks. You will still meet the same questions, with less vocabulary to think about them.

What the Vatican has actually proposed

The institutional Catholic position on AI — codified in the Rome Call for AI Ethics (2020), the Pope's 2024 World Day of Peace message, and a sequence of papal addresses — is not a refusal of the technology. It is a framework of constraints: AI must be transparent enough to be questioned, inclusive enough not to entrench existing inequalities, accountable to identifiable humans, and bounded by the dignity of the person it touches.

None of those constraints are uniquely Catholic. What is distinctive is the moral weight the institution can put behind them. When the Vatican signs a position, it commits the institution to it across the global Catholic communion. That is a different kind of binding than a Silicon Valley pledge.

The case for a religious seat at the AI ethics table

• –Most users of AI systems live inside moral frameworks shaped by faith traditions, not academic philosophy. Excluding those frameworks is excluding the people who use the tools.
• –Religious traditions have multi-century experience with questions about life, death, identity, and personhood. Engineering does not.
• –Religious authorities can mobilise consent. A policy that has the buy-in of major faith communities lands differently from one that does not.
• –Faith-based ethics often privileges constraints over optimisation, which is a useful counterweight in fields dominated by performance metrics.

The case against

• –No single faith speaks for all users. A formal religious seat at the AI ethics table raises the question of whose faith, immediately.
• –Religious authority on moral questions has historically been used to suppress minorities. Adding it to AI governance risks recreating that pattern in a new domain.
• –Several religious positions on AI are difficult to reconcile with the actual decision space of contemporary AI development — they describe an ideal that ignores constraints engineers cannot remove.
• –Secular ethics frameworks (Kantian, utilitarian, virtue-ethical) already incorporate the insights religious traditions also reach. The wheel does not need re-inviting.

The middle position most institutions are settling on

In practice, most large AI labs and government bodies are arriving at the same uneasy compromise: consult religious voices as one of several stakeholder groups, but do not give any of them formal veto power. The Rome Call for AI Ethics works this way. So does the Holy See's participation in the UN AI advisory body. The compromise satisfies neither the strong secularists nor the strong religious traditionalists — which is usually a sign that it is approximately right.

What this looks like in concrete dilemmas

The dilemmas in this cluster make the abstract question concrete. Should a tech company formally commit to a religious leader's AI framework? Should AI ethics boards include religious voices by design? Can an app that generates personalised prayers be a real religious practice? These are not abstract. Each is being decided in 2026 by people whose names you will not see in the press release.

Educational content. The Vatican's positions on AI are paraphrased from published documents (Rome Call for AI Ethics, 2024 World Day of Peace message, papal addresses). Positions attributed to other faith traditions are general summaries; specific authorities within each tradition hold a range of views. The SplitVote scenarios are hypothetical and intended for moral reflection, not policy recommendations.

Frequently asked questions

What is the Rome Call for AI Ethics?

A document signed at the Vatican in February 2020, originally by IBM, Microsoft, the FAO, and the Pontifical Academy for Life. It commits signatories to six principles — transparency, inclusion, responsibility, impartiality, reliability, and security and privacy — when developing and deploying AI. Other tech firms and several governments have signed in the years since.

Has the Pope written about AI?

Yes. Pope Francis dedicated his 2024 World Day of Peace message to AI, addressed the G7 in 2024 on the same topic, and the Vatican has continued to publish on AI governance through 2026. The institutional position consistently frames AI as not morally neutral and lists ethical limits — without rejecting the technology itself.

Do other religions have positions on AI?

Yes. The Buddhist community has emphasised mindfulness and right intention in AI design. Several Islamic scholars have published on AI through the lens of maslaha (public interest) and have raised questions about autonomous weapons. Jewish bioethicists have engaged closely with AI in healthcare and end-of-life decisions. The frameworks differ; the seriousness does not.

Why would a tech company listen to religious leaders?

Three pragmatic reasons beyond personal belief: religious institutions represent billions of users whose adoption decisions matter; faith traditions encode moral reasoning that survived centuries of scrutiny; and AI governance keeps colliding with end-of-life, identity, and dignity questions where secular philosophy and religious philosophy have both done serious work.

Is religious input on AI policy a violation of separation of church and state?

Not in most legal systems. Religious leaders advocating for policy positions is constitutionally protected speech. The line is at formal religious authority deciding the policy — which is different from religious voices being represented in the room.