EkklisiaAI for the ΑΩ

What monasteries are doing with AI: eight teams building tools for the work nobody sees

15 November 2026 · 7 min read

Eight monastic teams in Ukraine are building AI tools for the work nobody sees — with the guardrails written first.

Most of this site's recent articles have been about policies and guidelines — what churches should and should not do with AI. This week, eight monastic teams in Ukraine started from the other end: not a policy, but a project. On 10 November, eight teams from monastic and religious communities of the Ukrainian Greek Catholic and Roman Catholic Churches began practical training in the use of artificial intelligence, in an “AI Accelerator” programme run by the Platform for Social Change. The point, as the organisers put it, is to “streamline their internal processes and free up time for serving others.”

That sentence is worth sitting with, because it gets the order right. The tools are not the ministry. The tools are for the work around the ministry — the coordinating, the reporting, the searching, the organising — so that the time they save goes back into serving people.

What they are actually building

The teams will work on solutions for four kinds of everyday work: coordinating projects, preparing reports, searching for information in internal documents, and organising educational and charitable activities. Nothing here is about preaching, liturgy or pastoral care. It is the unglamorous middle of church life — the part that eats evenings and weekends and never appears in a sermon — and it is exactly where a small church feels the pinch most.

The training itself is practical, not theoretical: six practical sessions, independent work, and individual consultations. Lecturer Oleksiy Minakov will help participants master AI tools and develop the first working versions of their own solutions; mentor Vladyslav Prysyazhnyuk will guide the teams as they test and refine them. The shape matters — a teacher, a mentor, and teams building their own tools for their own work, rather than being handed a generic product and told to find a use for it.

This is not their first step, either. In September and October the Platform for Social Change ran three free online webinars for monastic communities on the responsible use of AI: on 24 September, digital communication — preparing announcements, responding to questions, handling feedback; on 8 October, developing an AI assistant based on a community's own internal materials; and on 22 October, working with donor information, reporting, and communicating the results of their activities. The ethical component of those sessions was led by Fr. Mykhailo Melnyk, a priest of the Kyiv Archeparchy of the Ukrainian Greek Catholic Church and founder of the Platform for Social Change. So the Accelerator lands on ground already prepared: people who have thought about the ethics first, and are now building.

The guardrails came first

That is the part of this story a local church should steal outright — because the Ukrainians wrote their guardrails before they built anything, and wrote them in human terms, not technical ones.

“Special attention will be given to the responsible use of technology in accordance with the principles of the Church's social teaching,” the report says. “The organizers emphasize the importance of protecting confidential information and ensuring human oversight. Personal, pastoral, medical, and other confidential data will not be shared with AI tools. Spiritual guidance, crisis response, and managerial decisions will remain the responsibility of people.”

Read that list again. It names what never goes into the tool — personal, pastoral, medical and other confidential data — and what never leaves human hands — spiritual guidance, crisis response, managerial decisions. That is a policy in two sentences, and it is stricter and clearer than most church AI policies written in committee. It also maps neatly onto the lines this site has drawn before: the Swiss guidelines' rule that no pastoral-confidential information goes into AI, and Austria's rule that AI in preaching, liturgy and pastoral care may only ever support, never replace. The monasteries are not contradicting those frameworks. They are showing what the frameworks look like when a community lives inside them and still gets on with the work.

There is a quiet honesty in the donor-information webinar, too. A community that handles donors' details and reports on its charitable work cannot afford to be casual about where that information goes. The October session took that head-on — working with donor information, reporting, and communicating results — rather than treating data protection as an afterthought to the exciting part.

Judge it by the results, at the end

The other thing worth borrowing is how the programme ends. At the end of the training, the teams will present the tools they have developed and the results of their implementation — including whether these tools have helped save time and improve the quality of their work. “This will give the communities a basis for deciding how to use and further develop their projects.”

No claim is made in advance that the tools will work. The verdict comes after the building, on evidence: did it save time, did it improve the work? If yes, keep going; if not, say so. That is a healthier test than either enthusiasm (“AI will transform our ministry”) or suspicion (“AI has no place here”). It is also the test this site's policy template suggests for a local church: try one small, bounded use, name who is responsible, and review it honestly before adopting anything else.

That ending also gives this story a natural sequel. When the teams present their results, we will know which of the four kinds of work — coordinating, reporting, searching, organising — actually benefited, and how. A follow-up piece then would not be speculation. It would be a report on what eight real communities found, in their own words.

What to borrow

So here is the practical takeaway, in four borrowable moves — each of them small enough for a congregation to try this month.

First, start with the work nobody sees. Not the sermon, not the pastoral visit — the reports, the rotas, the searching through old documents, the coordinating of a charity drive. That is where AI's help is least risky and most immediately felt, and where a bad result costs an afternoon, not a relationship.

Second, write the guardrails before you build anything, in two short lists: what never goes into the tool, and what never leaves human hands. The monasteries' version — no personal, pastoral or medical data in; spiritual guidance, crisis response and decisions out — is a good template to copy almost word for word, and to put in your own policy.

Third, build small and test it with a mentor, not alone. Six sessions, independent work, and someone to consult when the tool does something odd: that is a better model than one keen volunteer experimenting in private. If your church has no mentor, find one person who understands the tool and one who understands the data, and let them work together.

Fourth, decide on evidence, at a fixed point. Pick a date, look at whether the tool saved time and improved the work, and then decide — keep it, change it, or drop it. A decision made on results is one the whole congregation can trust, including the people who were sceptical at the start.

The monasteries are not showing us a finished answer. They are showing us a way of working: start small, guard what matters, build together, and judge by the fruit. A local church could do worse than copy the method — and a good deal worse than copy the guardrails.

Sources: Ukrainian Greek Catholic Church, “Monastic Communities of the Ukrainian Greek Catholic Church Begin Practical Training in the Use of Artificial Intelligence,” 13 November 2026 — training began 10 November; eight teams from monastic and religious communities of the Ukrainian Greek Catholic and Roman Catholic Churches; the AI Accelerator programme organised by the Platform for Social Change, with the aim “to help communities streamline their internal processes and free up time for serving others”; the four areas of work (coordinating projects, preparing reports, searching for information in internal documents, organising educational and charitable activities); six practical sessions, independent work and individual consultations; lecturer Oleksiy Minakov and mentor Vladyslav Prysyazhnyuk; the responsible-use passage quoted verbatim (confidential personal, pastoral and medical data not shared with AI tools; spiritual guidance, crisis response and managerial decisions remain with people); the final presentations, including whether the tools saved time and improved quality, as the basis for deciding further use; the three September–October webinars (24 September digital communication, 8 October an AI assistant from internal materials, 22 October donor information, reporting and communicating results) and Fr. Mykhailo Melnyk leading their ethical component. One limit, stated in the article: this is a single source, the UGCC's own report of the programme, and it gives no detail on which AI tools or models the teams are using, no end date for the training, and no results yet — the results are the point of the final presentations, so any follow-up must wait for those.

Related reading: Austria just wrote the first binding national church AI rules — and the first rule is about people, not technology and Five traditions, one voice: the Swiss churches just published the AI guidance the others are still writing — the two policy frameworks this piece's guardrails echo; A church AI policy in a weekend — the template with the named-owner step and the try-one-small-use test this piece points to. This guidance is general and educational, not legal advice; your church's own safeguarding, data-protection, and denominational policies always take precedence.