Ethics in practice

AI and human dignity: a Catholic Social Teaching framework

You do not need to be religious to use one of the oldest, most tested frameworks for putting people before technology. Catholic Social Teaching has spent more than a century working out how human dignity survives industrial change — and it turns out to be a remarkably practical lens for deciding where AI belongs in a business, and where it doesn’t.

By Alex CarpenterUnlocking TechnologyUpdated June 20269 min read

The one thing to take away

Every framework for “ethical AI” eventually has to answer one question: what is a person worth, and is that worth conditional on output? Catholic Social Teaching answers without flinching — human dignity is inherent and unconditional — and everything practical follows from there. Adopt the tools that serve that dignity; refuse the ones that quietly trade it away. That single commitment will steer more decisions correctly than any compliance checklist.

The central claim

AI should be adopted in the measure that it serves the human person and the common good — and resisted in the measure that it reduces people to data, concentrates power, or strips work of its dignity.

Four principles, applied to AI

Catholic Social Teaching rests on four permanent principles. They were written for an industrial age, but they read almost as if AI were in view. Here is each one against the concrete choices a small firm actually faces.

The dignity of the human person

The first principle is that every person has an inherent, non-negotiable worth that does not come from their output. AI pushes against this whenever it treats people as data to be optimised — ranking staff, scoring customers, nudging behaviour. Antiqua et Nova, the Vatican’s 2025 note on AI, is blunt that machine “intelligence” is not human intelligence and must never become the measure of a person. The test for any system is simple: does it treat the people it touches as ends in themselves, or as means to a number?

In practiceBefore deploying, ask whose dignity the system could quietly erode — staff, customers, applicants — and design that risk out.

The common good

Technology is meant to serve the flourishing of all, not concentrate advantage in a few hands. For a small firm this is practical, not abstract: an AI tool that lifts your whole team’s work, improves what customers receive and strengthens your community’s economy serves the common good. One that cuts corners on safety, truth or fair dealing to win a short-term edge does not — and tends to cost you trust later.

In practiceJudge each use by who benefits and who bears the cost. Favour gains that are shared with staff and customers, not extracted from them.

Subsidiarity

Decisions should be made at the most local, human level capable of making them well — and higher levels exist to support that, not to swallow it. Applied to AI, subsidiarity is the principled case for keeping a human in the loop: automate the mechanical, but leave judgement with the person closest to the work and its consequences. Centralising every decision into a model is exactly the failure subsidiarity warns against.

In practiceMap which decisions a model may inform and which a named person must own. Push authority down, not up into the system.

Solidarity

We are responsible for one another, with a particular duty to those most easily left behind. AI can widen gaps — between staff who are fluent and those who are not, between firms with capital and those without. Solidarity asks you to adopt in a way that brings your whole team along: training, redeployment rather than discard, and care for the customer who cannot or will not deal with a bot.

In practicePlan for the people the change is hardest on first. Budget for training and a human alternative, not just the licence.

The dignity of work

The tradition’s teaching on work is unusually relevant here. It begins with Rerum Novarum in 1891 — Pope Leo XIII’s response to the upheaval of the industrial revolution — and runs to the present: Pope Leo XIV took his name in 2025 with that history explicitly in mind, naming artificial intelligence as the new challenge to human dignity, justice and labour. Three commitments carry straight into an AI rollout.

Work has dignity in itself, not only in its product. A tool that hollows out the meaningful parts of a job while leaving the drudgery is a poor trade, even when it is cheaper.
The worker comes before the system. The point of automating a task is to free a person for better work, not to free the business of the person.
People have a right to participate in decisions that reshape their work. Imposing AI on a team is both an ethical failure and, the evidence shows, the surest way to make it fail in practice.

This is the same conclusion the research reaches from the other direction. Across the studies, the realistic and durable mode for AI in a smaller firm is augmentation, not replacement — which is why the ethics and the evidence point the same way. We make that practical case here.

What the Church actually said about AI

In January 2025 the Vatican published Antiqua et Nova, a detailed note on artificial intelligence and human intelligence. It is measured, not alarmist, and three of its points are worth holding onto whatever your beliefs:

AI is a tool. Its outputs can be impressive, but it does not understand, intend or take responsibility — and calling it “intelligence” should not blur that line.
Responsibility stays human. A model can inform a decision; it can never be the one held to account for it. Someone always has to answer for the outcome.
Judge AI by whether it serves truth, the common good and the dignity of work — not by novelty or efficiency alone.

A short test before you deploy

You can turn all of this into four questions to ask of any AI tool before it touches your business. If you cannot answer them, you are not ready to deploy it.

Dignity: Does this treat everyone it touches as a person, not a data point?
Common good: Are the benefits shared with staff and customers, or extracted from them?
Subsidiarity: Does a named human still own the decisions that matter?
Solidarity: Have we planned for the people this change is hardest on?

Put people first, without standing still

If you are weighing an AI tool and want to be sure it serves your team rather than hollowing it out, bring me the workflow or decision and I'll help you think it through — plainly, and honest about the trade-offs.

Get in touch

Keep reading

Augmentation, not automation — why AI realistically supports your team rather than replacing it — and how to roll it out that way.
A practitioner’s guide to AI for Australian SMEs — eight evidence-backed principles for where AI actually helps a firm your size.

About this piece. It reads four principles of Catholic Social Teaching — human dignity, the common good, subsidiarity and solidarity — alongside the Vatican’s 2025 note Antiqua et Nova and the current research on AI in small firms. It is offered as a practical ethical lens for anyone, of any belief, who wants technology to serve people. It is a framework for judgement, not a guarantee.