AI and the Church

The self-examination principle in Catholic tradition

Before applying the framework, the principle. Catholic teaching has consistently held that the Church must apply to itself the standards it asks of others. The principle goes by several names in the tradition: the unity of teaching and witness, the integrity of the Christian life, the rejection of hypocrisy. It is articulated across the Gospels, in the Pauline epistles, in the patristic tradition, in the great reform movements of every century, and in the documents of the Second Vatican Council.

The principle is not merely about appearances. It is about the substantive claim that the Christian message is the witness of a community that lives what it teaches. Christian teaching that is not embodied in Christian practice loses, in the tradition's own self-understanding, much of its capacity to call the wider world to anything. The Church that teaches about justice while practicing injustice teaches less effectively than the Church that practices what it teaches, however imperfectly. This is not a sociological claim. It is a theological one.

Applied to AI, the principle requires that Catholic institutions deploying AI examine themselves against the same framework they articulate for others. The questions Catholic teaching has been asking of public authorities, AI companies, secular employers, and individual users are the same questions Catholic institutions should be asking of themselves. The answers should be at least as demanding internally as the standards the Church articulates externally.

This page is built on the assumption that Catholic readers want their institutions to live the framework, not just to teach it. The assumption may be optimistic. The page is written for those who share it.

AI in parish life

Begin with parishes, because parishes are where most Catholics encounter the institutional Church. The AI uses in parishes are typically modest, but they are real and they are spreading.

Administrative work. Bulletin production, email communication, scheduling, fundraising appeals, basic correspondence. AI tools can save staff time on these tasks. Catholic teaching has no general objection to administrative efficiency, and parishes that use AI for these purposes are not categorically problematic. The questions are about whether the administrative use begins to drift into pastoral or liturgical territory, and whether the parish staff retains the human judgment the work actually requires.

Bulletin reflections and pastoral writing. This is where the drift typically begins. A pastor who uses AI to draft a bulletin reflection is in different territory than a pastor who uses AI to schedule volunteer ministers. The bulletin reflection is the pastor's pastoral voice to the parish. AI-generated content represented as the pastor's reflection is a misrepresentation that the parishioners may not notice but that the framework names as problematic. The honest version is one of two: the pastor writes the reflection himself, however imperfectly, or the bulletin is clearly identified as reprinting or summarizing other authors.

Homily preparation. The homily occupies a particular place in Catholic worship. It is not merely instruction; it is the personal preaching of the Word of God by a particular ordained minister to a particular congregation. AI may serve a homilist the way commentaries and concordances have served homilists, as a starting point and research aid. AI should not be the source of the homily that is preached. The congregation has the right to receive a word the preacher has actually engaged with. The line is not always easy to draw, but it exists, and the homilist has the responsibility to keep on the right side of it.

Pastoral communications. The personal letter to a grieving family, the response to a parishioner in crisis, the email to a couple preparing for marriage. These are pastoral acts, not administrative ones. AI assistance here, even when convenient, removes the pastoral presence that the communication is supposed to embody. The parishioner reads the letter assuming it carries the pastor's actual engagement. If it does not, the trust is being misused.

Practical guidance for parishes. A simple parish AI policy goes a long way. The policy should identify which uses are clearly acceptable (administrative work, scheduling, basic correspondence), which require particular care (bulletin content, communications to parishioners), and which are not acceptable (homilies, pastoral letters, liturgical content). The policy should be made known to staff and reviewed annually. Most parishes that have one find it both clarifying and reassuring.

AI in diocesan governance and operations

Dioceses are now substantial institutional employers, with HR functions, payroll, communications, legal departments, education offices, and significant administrative complexity. AI is entering all of these.

Diocesan personnel and HR. Many dioceses now use AI tools in some part of their hiring process, particularly for screening applicants for school, parish, and Catholic Charities positions. The same Catholic framework that applies to secular employers applies here. AI screening tools should not be the sole basis for hiring decisions. Candidates should know that AI is being used. The patterns of decisions should be audited for bias. Human review should be meaningful, not cosmetic. Catholic teaching on the dignity of work applies to every employee a diocese hires, and the AI tools used in the hiring should reflect that.

Diocesan communications. AI-assisted communications are increasingly common in diocesan offices: press releases, social media content, internal memos, fundraising appeals. The general framework permits this with caution. The line to watch is between administrative communications and pastoral or magisterial communications. A diocesan announcement about a building project is in a different category than a diocesan statement on a moral question. AI assistance is more appropriate to the first than the second.

Pastoral planning and parish allocation. Some dioceses are now using data analytics tools, sometimes with AI components, to inform pastoral planning decisions: parish mergers, school closures, priest assignments. These decisions have significant consequences for the communities affected. The Catholic principle of subsidiarity supports the meaningful participation of affected communities in such decisions. AI tools that make these decisions feel more objective should not be allowed to obscure the moral weight of what is being decided or to substitute for the consultation that the Catholic framework requires.

Legal and risk management. Diocesan legal offices, like other legal offices, are now using AI tools for document review, research, and case management. The general framework permits this. The questions are about confidentiality of communications, the protection of sensitive information (including information related to clergy abuse cases), and the maintenance of human professional judgment in the work. Diocesan general counsel should be tracking these questions deliberately.

Practical guidance for dioceses. Each diocese should have a written AI policy approved by the bishop, with appropriate input from canonical advisors, lay professionals, and ethics consultants. The policy should be made public to the extent appropriate, since transparency on these questions is itself part of Catholic accountability. The policy should be reviewed regularly. Bishops who delegate AI questions entirely to staff are likely to find themselves accountable for decisions they did not adequately oversee.

Catholic schools and universities

Catholic educational institutions face the most acute AI questions in their daily work. Students bring AI tools into every assignment. Teachers face decisions about how to teach in an AI-saturated environment. Administrators make procurement decisions about AI-driven curriculum software, plagiarism detection, scheduling, and enrollment management.

Student AI use. The question of how students should use AI in their education is the most pressing one in Catholic schools right now. The categorical positions, "ban AI entirely" and "permit AI freely," both fail. The framework Catholic teaching would support is more nuanced. AI as a research and tutoring aid that supports the student's own intellectual work is acceptable. AI generating work the student then submits as their own is academically and morally problematic. The middle cases are where the school's policy has to do its work. The school's policy should be explicit, taught to students as part of their formation, and applied consistently. The policy should distinguish among subjects, grade levels, and types of assignments.

Teacher AI use. Teachers using AI to grade assignments, generate feedback, write reports, and produce lesson materials raise questions parallel to the homily question. The teacher's personal engagement with the student's work is part of what the relationship is for. AI that mediates that engagement too heavily compromises it. The school policy should address teacher AI use as deliberately as it addresses student use.

AI plagiarism detection. Catholic schools using AI tools to detect student AI use should be aware of the documented bias of these tools, particularly against non-native English speakers. False accusations are not a minor administrative inconvenience; they damage students. Schools deploying these tools should require human review of any case the tool flags, should communicate clearly with students about what is being used, and should be willing to revise or abandon tools that produce biased outcomes.

Admissions and enrollment. Some Catholic schools and universities use AI tools in admissions decisions. The same standards that apply to other employers and educational institutions apply here. The Catholic framework would press additionally on whether the tools are reproducing patterns of exclusion that Catholic teaching has identified as problematic, particularly toward students from lower-income or minority backgrounds.

Catholic universities' research role. Catholic universities have a particular vocation in this moment to produce the scholarship and the educated public discussion that secular institutions are not always positioned to produce. The Catholic intellectual tradition has resources for thinking about AI that secular bioethics programs do not have. The universities should use these resources publicly. Internal AI policy should make space for faculty engaged in this work and should not pressure them toward uncritical adoption of tools that are themselves the object of their inquiry.

Catholic hospitals and healthcare systems

Catholic healthcare is one of the largest sectors of Catholic institutional life, employing millions of workers and serving tens of millions of patients globally. AI in healthcare is entering at multiple levels: clinical decision support, diagnostic imaging, administrative scheduling, electronic health records, patient communication, billing. The Catholic identity of the institutions raises particular standards.

Clinical AI tools. AI in clinical decision-making, particularly in diagnostic imaging and treatment recommendations, is now routine in major healthcare systems including Catholic ones. The Catholic framework permits the use of such tools, with conditions. The physician retains responsibility for the diagnosis and treatment plan. The AI's recommendations should be auditable in some meaningful sense. The patient should know AI is involved. The tools should be evaluated for bias in their training and deployment. None of this is unique to Catholic healthcare, but Catholic identity makes it more important to do well, not less.

End-of-life and palliative care. Catholic teaching on end-of-life care is among the most developed in any tradition, with substantial doctrine on the dignity of the dying, the use of ordinary and extraordinary means, and the prohibition on euthanasia and physician-assisted suicide. AI tools that shape end-of-life decisions, including AI mortality predictions that influence treatment intensity, should be approached with particular care. Catholic healthcare systems should ensure that the AI does not function as a tool for cost-driven withdrawal of care, and that the patient's dignity remains central regardless of what the algorithm projects.

Administrative deployment. Catholic hospitals are large employers, and the AI deployed in HR, scheduling, and worker management is subject to the same Catholic framework on the dignity of work that applies to other Catholic employers. Catholic healthcare workers should know what tools are operating on them and should have access to the appeal mechanisms Catholic teaching insists on.

Patient communication. AI chatbots used in patient communication should be clearly identified as such. Patients have a right to know when they are interacting with an AI and when they are interacting with a person. Catholic teaching's insistence on the integrity of human relationships extends to the doctor-patient and nurse-patient relationships, which are foundational to healthcare. AI should not be allowed to substitute for these without clear patient consent.

Pastoral care in the hospital. One area where the Catholic line should be especially clear: pastoral care, chaplaincy, and the sacramental life of the patient cannot be mediated through AI. The dying patient who asks for a priest needs a priest, not an AI chatbot trained on theological texts. The Catholic hospital's chaplaincy department should be unambiguous about this, and should resist any institutional pressure to "scale" pastoral care through technological substitutes.

Religious orders and formation

Religious orders, monasteries, and other forms of consecrated life are now considering AI tools in formation, vocations work, and community life. The Catholic framework here has a particular texture because the consecrated life is structured around practices that AI substitutes can readily undermine.

AI in formation. The formation of novices and seminarians is not transmission of information. It is the formation of persons in a particular way of life through specific practices: prayer, community life, study under qualified directors, spiritual direction, examination of conscience. AI assistance in any of these would substitute for what the formation is supposed to do. A novice who uses an AI to draft her spiritual reflection assignments is not being formed by the practice of writing those reflections; she is producing artifacts that look like formation without undergoing it. Formation directors should be alert to this.

Spiritual direction. AI-driven spiritual direction tools are now available commercially. The Catholic teaching on this is direct. Spiritual direction is a personal accompaniment that depends on the genuine human presence and pastoral judgment of the director. AI substitutes for this are not adequate to what spiritual direction is supposed to be, regardless of how sophisticated the tools become. Religious orders should be unambiguous in directing members toward genuine spiritual direction by qualified human directors.

Vocations work. Many religious orders use digital tools, sometimes including AI elements, in vocations outreach. This is generally permissible, with the same caveats that apply to other AI-assisted communications: clarity about what the technology is doing, appropriate human follow-up, attention to the genuine encounter the discernment process requires.

Community life. Religious community life is built on practices that AI can erode if introduced thoughtlessly: shared meals, common prayer, recreation together, ongoing conversation. The introduction of AI tools into community life should be deliberate, with attention to what the technology is doing to the practices that constitute the community. Religious orders are unusually well positioned to think about this carefully, because they have a long tradition of attention to the conditions under which community life flourishes.

Catholic Charities and Catholic social services

Catholic Charities, the Society of Saint Vincent de Paul, Catholic Relief Services, and the broader network of Catholic social service organizations are major deployers of AI tools, often without the public attention that for-profit AI deployment receives.

Beneficiary screening. Catholic charities providing services to vulnerable populations face the same algorithmic decision-making concerns that the AI and the Poor page addresses for secular agencies. AI tools used in determining eligibility, allocating resources, or prioritizing cases should be evaluated against the standards Catholic teaching insists on. The people being served are precisely the people Catholic teaching identifies as the focus of the preferential option for the poor.

Case management AI. AI-assisted case management is becoming common in Catholic social service agencies. The Catholic framework permits this, with caveats. The case worker retains professional judgment about the client. The AI is a tool, not a substitute. The client knows AI is involved if their case is being shaped by it. Confidentiality is protected. None of this is exotic; it is what Catholic social work has always required of its tools.

Grant writing and fundraising. AI assistance in grant writing and donor communications is widespread in Catholic charities. This is generally acceptable, with attention to representations made to funders. A grant proposal generated by AI that claims to reflect the agency's distinctive Catholic approach should actually reflect what the agency does, not what the AI thinks a Catholic agency should do.

Mission integrity. Catholic social service agencies are not generic non-profits. Their Catholic identity makes specific claims about how they operate. AI tools that would homogenize their operations toward generic best practices may compromise that identity. The agencies should be willing to refuse tools whose use would erode their distinctive mission, even when the tools would improve metrics.

The Vatican and the Roman Curia

The Vatican itself is now using AI tools, experimentally and increasingly. Translation services for papal documents, summarization of long inputs, analysis of large document collections, and other administrative uses are documented. The scope is not as broad as in some commercial institutions, but it is real.

The Catholic framework's application to the Vatican is the highest-stakes application, because the Vatican's institutional behavior is read globally as a signal of what the Church actually means by its teaching. AI use in the Curia that does not reflect the framework Magnifica Humanitas articulates would be a public failure of integrity. AI use that does reflect the framework would be a model the wider Church could learn from.

The specific decisions are made by the dicasteries, the Secretariat of State, the various Vatican offices, and increasingly by AI-focused bodies within the Holy See. Public transparency about what is being used and why would itself be a contribution. The Vatican's own AI policy, when it is articulated, should be made public.

One specific area deserves attention. Vatican communications, including papal messages and documents, depend for their authority on being understood as the Pope's actual words and teaching. AI assistance in producing such documents would compromise that authority in ways the Catholic faithful and the wider world have not consented to. The strong default here should be against AI drafting of magisterial content, even when AI is used in supporting research.

A particular caution: liturgy and the sacraments

One area requires standalone treatment because the Catholic standards are unusually clear and the temptation to drift is unusually strong.

Catholic liturgy is the public worship of the Church, in which the people of God offer themselves to God through Christ. Sacred art, liturgical music, the prayers of the Mass, the homily, and the sacraments themselves are not generic religious content. They are constitutive of the Church's worship, and the standards governing them have been developed over centuries with great care.

AI in liturgical contexts should be approached with strong defaults toward refusal. AI-generated sacred art for parish use, AI-composed liturgical music, AI-written prayers used at Mass, AI-illustrated children's catechism materials for sacramental preparation, all sit in a category where the Catholic tradition has been clear for centuries that human creative offering is integral to what the work is.

The sacraments themselves cannot be mediated through AI. Baptism, confirmation, Eucharist, reconciliation, anointing of the sick, marriage, and holy orders all require the personal action of the minister and the participation of the recipient. AI cannot baptize, cannot hear confession, cannot anoint, cannot ordain. This is not a question that admits of technological evolution. The sacramental life of the Church is what it is, and the Church's identity as the body of Christ depends on its continuing to be what it is.

Catholic institutions deploying AI elsewhere should keep these lines bright. The administrative AI in the parish office is one thing. The AI-generated hymn in the Sunday Mass is another. The Catholic standards are different in different domains, and the institutions should know the difference.

What this asks of Church leaders

The framework is most useful when it produces specific responsibilities for specific people. For Church leaders, the responsibilities are concrete.

Make AI policy a leadership concern, not a delegated one. Bishops, religious superiors, university presidents, hospital system CEOs, and other Catholic institutional leaders should engage AI policy directly. Delegating the question entirely to IT departments or to outside consultants is not adequate to what the framework requires. The decisions are moral and theological, not just technical.

Be willing to refuse adoption. The pressure on Catholic institutions to adopt AI tools comes from many directions: efficiency, cost, peer institutions, vendor marketing, board members. The Catholic framework permits adoption with conditions, and it sometimes requires refusal. Leaders should be willing to make the refusals when the framework requires them, even when the institutional pressure runs the other way.

Be transparent. Catholic institutions deploying AI owe their members, employees, beneficiaries, and broader publics a clear account of what is being used, why, and with what safeguards. Internal opacity on AI questions is corrosive of the trust Catholic institutions depend on. The default should be public disclosure, with appropriate exceptions for genuinely confidential information.

Form your people. The clergy, lay ministers, teachers, healthcare workers, and other Catholic institutional employees need formation to engage AI questions in their daily work. The formation cannot be left to chance or to whatever the employees pick up from the broader culture. Institutional leaders should ensure that their people are equipped to apply the Catholic framework in their actual work.

Be accountable. Leaders should expect to be asked, by their faithful, their employees, their beneficiaries, and their broader publics, how they are using AI and why. The willingness to answer these questions honestly, including in the cases where the answers are unflattering, is part of what Catholic leadership in the AI era requires.

The encyclical releasing on May 25 is going to make the framework public and unmistakable. Catholic institutional leaders who have not engaged the framework will find themselves accountable for institutional practices they cannot adequately defend. The work begins now, not after the document arrives.