Lecture 15-1: The Pastoral Ethics of Administration, Data, and Privacy

Proverbs 11:13 — "One who brings gossip betrays a confidence, but one who is of a trustworthy spirit is one who keeps a secret." (World English Bible)

1 Corinthians 14:40 — "Let all things be done decently and in order." (WEB)

Acts 6:1-7 — "Now in those days, when the number of the disciples was multiplying, a complaint arose from the Hellenists against the Hebrews, because their widows were neglected in the daily service. The twelve summoned the multitude of the disciples and said, 'It is not appropriate for us to forsake the word of God and serve tables. Therefore select from among you, brothers, seven men of good report, full of the Holy Spirit and of wisdom, whom we may appoint over this business. But we will continue steadfastly in prayer and in the ministry of the word.' These words pleased the whole multitude. They chose Stephen, a man full of faith and of the Holy Spirit, Philip, Prochorus, Nicanor, Timon, Parmenas, and Nicolaus, a proselyte of Antioch; whom they set before the apostles. When they had prayed, they laid their hands on them. The word of God increased and the number of the disciples multiplied in Jerusalem exceedingly. A great company of the priests were obedient to the faith." (WEB)

3.1 Three places AI is already inside the church office

Start with what is actually happening, not a hypothetical. A growing number of church management platforms — the software that tracks membership, attendance, giving, and small-group participation — now market AI-driven analytics as a feature, not an add-on. Three uses are already common enough to name specifically.

First, member-risk flagging. A platform that logs attendance and giving over time can run a model that flags members whose pattern has changed — fewer Sundays attended, giving reduced or stopped — as candidates for pastoral follow-up. The pitch is reasonable on its face: busy staff cannot track hundreds of households by memory, and a system that surfaces a quietly drifting member before they leave altogether sounds like exactly the kind of "no one falls through the cracks" tool a caring church wants. The risk is what the flag does once it exists. A list generated by a giving-and-attendance algorithm is, definitionally, a list built from financial and behavioral surveillance of the congregation, generated and stored by a vendor, and it will be read differently by whoever receives it than a shepherd's own accumulated, relational sense of who has gone quiet. It converts a pastoral judgment call — who needs a call this week — into a data output that looks more authoritative than it is, and it does so using giving records, which carry a particular weight of confidentiality in most congregational traditions precisely because they are so easily misused to rank members by financial usefulness.

Second, AI-assisted screening of prayer requests or counseling intake forms. A platform that categorizes, summarizes, or routes prayer requests using a language model is processing some of the most sensitive disclosures a person makes to a church — a marriage in crisis, a diagnosis not yet shared publicly, a struggle with addiction or abuse. Summarization sounds like a convenience (staff read a two-line synopsis instead of a paragraph), but a summary is also an interpretation, produced by a system with no relationship to the person and no accountability for getting the interpretation wrong, and the underlying text has now been sent to a third-party server to be summarized at all.

Third, and underlying both of the above: third-party data processing itself. Almost no church runs its own AI infrastructure. A "smart" feature in church management software means the vendor — and very often a further sub-processor the vendor contracts with to provide the actual model — receives the data to process it. The church's privacy posture is therefore only as strong as a vendor's terms of service, which the person filling out an online intake form has certainly not read, and which may permit uses (model training on submitted data, indefinite retention, disclosure under a change of ownership) that the church itself would never knowingly agree to if the question were put to it directly rather than buried in a software agreement.

3.2 Why intention is not the test

A natural pastoral response to all three concerns is some version of: no one intends harm; the goal is care, not surveillance. That is almost certainly true as a statement of motive, and it is not a sufficient ethical answer, for two reasons.

First, structural risk does not require ill intent to be real. A vendor's data breach, a training-data leak that surfaces a sensitive detail in an unrelated context, a sub-processor's own security failure — none of these require the church, or even the vendor, to have wished any harm. Proverbs 11:13 draws exactly this line: the contrast is not between a malicious talebearer and an innocent bystander, but between "one who goes about as a talebearer" and "one who is of a trustworthy spirit," where trustworthiness is defined by concealing a matter, not merely by not intending to expose it. A congregation that routes sensitive disclosures through a system it has not vetted for exactly this kind of exposure has not thereby become a talebearer, but it has stopped being a reliable keeper of the matter in the sense Proverbs commends — reliability, on this text, is an active discipline, not a default state maintained by good will alone.

Second, the asymmetry of information matters morally. The member submitting a prayer request or filling out a counseling intake form is doing so inside a relationship of trust with a pastor or a church, not with a software vendor whose name they may never see. They have not consented to the vendor's data practices in any meaningful sense — clicking through a form is not informed consent to a data-processing chain the person cannot see and was never shown. A church that adopts a tool without examining what it actually does with sensitive data has, in effect, extended its own pastoral relationship of trust to a third party the member never agreed to trust. This is precisely the kind of quiet, structural decision the chapter's opening principle names: nothing was voted on, no one meant to expose anyone, and a real exposure now exists anyway.

3.3 Acts 6 as the positive model — delegation without loss of accountability

It would be a mistake to let this lecture read as pure caution, because the New Testament's own account of a church's first administrative crisis is not a caution story — it is a story about administration done well, and it has direct bearing on how a congregation should think about delegating tasks, including to a technological tool.

The situation in Acts 6:1 is a real operational failure: the daily distribution to widows was being administered unevenly, along an ethnic fault line (Hellenist widows against Hebrew widows) that the text does not flinch from naming. The apostles' response is instructive in what it does and does not do. It does not simply absorb the extra work themselves, and it does not simply hand the task to whoever is available without qualification. It creates a defined role — "over this business" (Acts 6:3) — filled by named, qualified, publicly accountable people: "seven men of good report, full of the Holy Spirit and of wisdom." The apostles lay hands on them; the appointment is public, prayed over, and attached to specific persons who can be held responsible for how the distribution is actually run.

This is the template a congregation should bring to any AI-assisted administrative tool, and it reframes the question productively. The issue was never whether administration may be delegated or made more efficient — Acts 6 delegates a whole operational function and the text calls the result good ("the word of God increased," 6:7). The issue is whether delegation is done with accountability (named people, a defined scope, a standard the people meet) or instead of accountability (a tool adopted because it is convenient, with no one able to say who is responsible for what it does with the data it touches, and no defined boundary on its use). An AI-driven flagging feature or intake-screening tool is not disqualified by being a tool rather than a person — Acts 6 already shows this is not the church's normal instinct — but it has not yet cleared the bar Acts 6 actually sets: a defined scope, and a named person answerable for it. 1 Corinthians 14:40's "let all things be done decently and in order" names the same standard from the other direction — not order as bureaucratic tidiness for its own sake, but as the condition under which delegated ministry, whether to seven qualified men or to a software feature, stays trustworthy rather than becoming a place where no one is quite in charge of what happens to a member's most sensitive disclosure.