Equipping Congregational Leaders for Ethical AI Engagement: Lessons from the ABHMS Grant Cohort

What the ABHMS Cohort Actually Is

The program is built in three tiers, and the order matters as much as the content. The first tier is a broad foundation open to more than fifty participants, delivered through bi-weekly hybrid sessions that pair guest speakers and case studies with explicit theological analysis. This is where leaders learn what AI is, how it works in plain terms, and what ethical questions it raises before they ever touch a tool in earnest. The second tier, "Skill Building," narrows to graduates of the first and moves into practical workshops on the responsible use of specific platforms. The third tier, "Innovation," takes a small group of advanced participants and asks them to actually build something useful for ministry, with results presented at a conference and folded into an ongoing ethics think tank.

Two structural choices stand out immediately. The program is tuition-free and delivered in a hybrid format specifically to lower barriers, so that a bivocational pastor in a rural church has the same access as a staff member at a large urban congregation. And it is led from inside the organization by someone with both technical and pastoral standing, Rev. Saeed Richardson, who serves as ABHMS Chief Technology and Information Security Officer. His framing is that congregational leaders need to be prepared to navigate emerging technologies and integrate them into ministry through frameworks rooted in faith, justice, compassion, and ethics. That sentence is doing real work: it names ethics and justice as the starting point, not a disclaimer bolted onto a software demo.

Lesson Two: Build a Progression, Not an Event

A single training session, no matter how good, produces a brief spike of interest and almost no lasting change. People forget, momentum fades, and the handful who were already inclined to experiment carry on while everyone else returns to old habits. The tiered cohort model solves this by treating AI literacy as a pathway with distinct stages and increasing commitment. The wide first tier creates broad exposure across more than fifty leaders. The narrower second tier rewards those who want to go deeper with hands-on skill. The small third tier concentrates resources on the few who will build durable, reusable applications for the wider community.

This funnel does something subtle and valuable. It lets a program serve very different levels of readiness at once without watering down any of them. The curious skeptic gets a foundation that respects their caution. The eager adopter gets practical skill without skipping the ethics. And the natural innovator gets a mandate and a small peer group to build something real. Crucially, the people who advance become a distributed bench of capable leaders rather than a single overworked tech champion, which is how knowledge actually spreads through an organization. We explore that dynamic of cultivating internal advocates in our guide to building AI champions.

The progression also creates a renewable structure rather than a finished product. Because the third tier feeds an ongoing ethics think tank, the cohort does not end when the grant does; it leaves behind a standing capacity to keep asking new questions as the technology changes. That is the difference between training people once and building an institution that keeps learning. For organizations thinking past the first workshop, our piece on building continuous AI learning pathways covers how to sustain momentum after the initial enthusiasm fades.

Lesson Four: Design for Access from the Start

It would have been easy to run this program as a paid, in-person intensive that drew the already-resourced congregations who could send a staff member for a week. The cohort was deliberately built the other way. It is tuition-free, removing the cost barrier that keeps smaller and rural congregations out of professional development. It is hybrid, removing the travel barrier that excludes bivocational leaders who cannot leave a second job. And it leans on existing digital infrastructure rather than requiring participants to buy new systems. These are not incidental details; they are an equity strategy, and they reflect the conviction that AI literacy should not become one more advantage that flows only to the largest institutions.

The wider danger here is real. As AI capability becomes a marker of organizational effectiveness, the communities that cannot afford training risk falling further behind precisely as the technology becomes more consequential. A program that prices in access from the first design decision pushes against that drift. For faith-based and other mission-driven organizations, this is also a theological and values commitment, not just a logistical one: the justice frame that ABHMS names explicitly means asking who gets left out of the room and then redesigning the room so they do not. Any nonprofit building AI training should treat access as a design constraint rather than a nice-to-have, and our look at the broader delivery principles for nonprofit AI training develops that point further.

The Boundaries the Cohort Keeps in View

A program that teaches AI well also teaches where AI does not belong, and the strength of an ethics-first design is that it keeps those boundaries central rather than leaving them to chance. For congregational leaders, the clearest lines fall around the relational and spiritual heart of ministry. Pastoral care, prayer, confession, and the moments of presence that define a faith community are not tasks to be optimized; they are the point. A tool that drafts an announcement is helpful. A tool that drafts a condolence to a grieving family, sent as though it were personal, betrays the trust the relationship was built on. The cohort's framing keeps that distinction in front of every participant.

The same discipline applies to truth and transparency. Because language models produce fluent, confident text regardless of accuracy, anything that carries doctrinal or factual weight has to be verified by a qualified human before it represents the community. And because trust collapses when people discover rather than are told that something was machine-generated, disclosure has to be a habit, not an afterthought. Our guide to AI and pastoral care works through exactly where faith communities should hold the line, and our overview of how denominational bodies are writing AI policies shows how those boundaries are being formalized so individual leaders are not left to improvise alone.