Five traditions, one voice: the Swiss churches just published the AI guidance the others are still writing

Five traditions, one document: the first finished ecumenical AI framework is out, and a local church can borrow from it this weekend. Reading time: about 5 minutes.

The denominations piece had an honest limit built into it: none of the three bodies it covered had published finished guidance yet. The Methodists had voted for a task force, the Anglicans had adopted terms of reference, and the Catholic bishops had a framework to work from. All slow, careful work — and all still to come.

There is now a fourth story, and it is different in kind. Five Swiss church bodies — Protestant, Roman Catholic, Old Catholic and evangelical, together — have jointly published ten ethical guidelines on artificial intelligence. The document is called Humanity in the Digital Age: Ethical Guidelines for the Responsible Use of Artificial Intelligence: An Ecumenical Christian Perspective. It was released on 19 October and reported by the World Council of Churches on 11 November.

Five traditions, one document. “On this question, we as churches in Switzerland have one voice,” said Rita Famos, president of the Protestant Church in Switzerland. That is the news in one sentence — and it is also the challenge to the rest of us. If five traditions that disagree about plenty can agree on ten theses about AI, a single congregation can surely agree on one page.

What the guidelines actually say

The ten guidelines rest on a Christian understanding of the human person: people are beloved creatures made in the image of God, and their dignity comes from God's unconditional love, not from performance or usefulness — which places it outside anything a machine can achieve. Bishop Charles Morerod, president of the Swiss Bishops' Conference, put the connection bluntly: “Whoever speaks about artificial intelligence is always speaking about the human being as well.”

The first four theses are the theology, stated plainly: people differ from machines; relationships depend on embodied experience; personal formation cannot be handed off to technology; and technology does not write history. The remaining six are the ethics that follow: truth and transparency; protecting human participation in decision-making; the claim that people are more than their data; environmental responsibility in the digital sphere; limits on concentrated digital power; and shared public responsibility for the direction of AI.

The authors reject both the utopian promises and the doomsday predictions about AI. Both stories, they argue, treat technology as the driving force of history and reduce people to onlookers — and both distract from the present problems: algorithmic discrimination, concentration of power, environmental costs. The greater risk, they say, is not that machines become capable of too much, but that people voluntarily surrender their own freedom and responsibility — out of resignation, or simply for convenience.

That is a sharper warning than most church statements manage, and it lands closer to home than any talk of superintelligence. The danger for a local church is not a robot in the pulpit. It is a volunteer pasting a prayer list into a chatbot because it is easier, and nobody asking whether that was a good idea.

The rules for church practice — and how familiar they look

Much of the document is about how churches themselves should use AI, and readers of this site will recognise the shape of it. Where AI is used in preaching, education, media work, counselling or fundraising, its output should be checked against sources — quotations, factual claims and legal information should not be accepted without verification. Churches should disclose AI use in counselling, pastoral care, education and decision-making, and wherever it shapes personal communication or authorship. A named individual must always remain responsible.

On sermons, the warning is pointed: lessons and preaching produced efficiently by software, without being worked through and spiritually wrestled with by the preacher, cause church communication to lose depth and truthfulness. That is the same line this site's guardrails piece draws — AI preps, people disciple — now with five traditions behind it.

The data-protection section sets the firmest limits. Personal or sensitive data should not be entered into AI systems unless storage location, further use and retention periods are contractually settled. The guidelines rule out AI-based profiling of people seeking help, employees or volunteers, and fully automated decisions in sensitive areas. Information covered by pastoral confidentiality must not be entered into any AI system at all. And churches must never use closeness, trust or spiritual authority to manipulate people through AI tools.

If that list sounds familiar, it should: it is the live policy template on this site (A church AI policy in a weekend), written in denominational language. The Swiss document is, in effect, the theological backing for a one-page policy a congregation can write this weekend — and a reason to trust that policy once it is written.

Not a ban — and that matters

The document does not urge churches to avoid AI. It describes a dual task: monitoring and assessing the technology critically, while investing in expertise, creating space for responsible experimentation, and sharing applications that work.

Among the benefits it names are real-time translation, automated transcription and text-to-speech tools, which let people with illnesses, disabilities or language barriers take part in church and public life. That is not abstract for this site's readers: live translation in the service is the newest entry in the tools directory, and the Swiss guidelines are a reminder to ask the questions that entry raises — what happens to the audio, who stores it, who is responsible when it gets something wrong — rather than treating the technology as either a miracle or a menace.

The authors are equally honest about the other side. Some people find it easier to open up to a chatbot than to another person, and that can be a step toward later speaking with someone face to face. But platforms built to capture attention and simulate emotional closeness can replace human relationships rather than supplement them, and children, adolescents and vulnerable people need spaces where trust and belonging are not generated by algorithms.

Several sections look beyond Switzerland. The most powerful AI systems are controlled by a few large companies, mainly based in the United States and China, and churches, schools and government agencies are becoming dependent on proprietary systems whose terms and prices they cannot influence. People in poorer regions carry much of the cost — through raw material extraction under precarious conditions, low-paid and stressful data work, and exclusion from digital participation — while environmental costs, including electricity and cooling water for data centres and growing electronic waste, are shifted elsewhere and remain largely invisible to users in the Global North. Bishop Frank Bangerter of the Christian Catholic Church put the test simply: “AI must not benefit only those who already hold power anyway. Technical progress is only good if it serves everyone.”

The practical advice follows from that: churches should assess AI products for data protection, transparency, technical dependency and the ability to switch providers — before adopting them, not after.

One honest footnote, and what to do with this

Two things worth knowing before you borrow from the document. First, the authors describe it as provisional and subject to ongoing review — a first step, not a final word. It is linked to the Charta Oecumenica, the joint commitment of European Protestant, Orthodox and Catholic churches, whose revised edition, signed in Rome in November 2025, includes a section on new technologies for the first time. The Swiss text calls itself a first step in that direction.

Second, the Swiss bishops have confirmed that AI tools were used intentionally in producing the guidelines themselves — for initial research and analysis, and in the final editorial process — with an ecumenical working group discussing, revising and theologically developing the framework, and more than 200 comments incorporated in consultation. That is not a scandal; it is the guidelines' own disclosure rule, applied to themselves. A church that uses AI openly, with named humans responsible, is doing exactly what the document asks.

So here is the practical takeaway. If your church has not yet written its one-page AI policy, the Swiss guidelines give you something the denominations piece could not: a finished, ecumenical text to borrow the principles from. Read the ten theses — the full text is available in multiple languages at menschlichkeit-digital.ch — and then do the weekend version: name an owner, fill in the template, tell the congregation. The Swiss churches did the slow, careful work so that yours does not have to start from a blank page.

“Churches are not AI experts,” said Andi Bachmann-Roth, co-general secretary of the Swiss Evangelical Alliance. “But behind the technology lies the question of what a flourishing life is. To that, Christianity has offered a sustaining answer for two millennia.”

Start here: A church AI policy in a weekend →