Artificial Intelligence Policy

Delbarton’s Philosophy and Core Values Regarding Artificial Intelligence (AI)

Our Approach Flows from Delbarton’s Identity and Mission
Delbarton School approaches artificial intelligence (AI) first as a matter of human formation. Thus, our Mission Statement, Belief Statements, Profile of a Graduate, and Hallmarks of a Benedictine School are the lenses through which the School evaluates AI and its use in our educational program.

The Rule of St. Benedict and the Catholic Church’s social and educational teachings also inform this discernment. The University of Notre Dame’s DELTA framework (Dignity, Embodiment, Love, Transcendence, Agency) serves as a complementary practical tool for applying these commitments to emerging technologies. Learn more about Notre Dame's work on the ethics of AI here.

Our Commitment

Delbarton will use AI to help form young men as brothers and leaders who think critically, work diligently, act honorably, serve the common good, and use technology wisely in lifelong service to God and neighbor. We will ask not only what AI can do, but what habits, relationships, judgments, and persons its use helps us form in our students.

Benedictine Principles for AI Use

● Listening and Obedience: Technology should support attentive listening, critical reflection, responsible freedom, and responsiveness to truth, teachers, community, and conscience, not replace them.

● Work and Discipline: Intellectual labor has formative value. AI should not remove the skills of reading, writing, reasoning, practice, perseverance, or craftsmanship that an assignment is designed to foster in a student.

● Humility: Students and adults must recognize the limits of their own knowledge and the limits of AI systems, including their capacity to generate errors, bias, or fabricated information.

● Community and Stability: Technology should strengthen, not displace, the human relationships through which learning and formation occur.

● Hospitality and Love of Neighbor: AI should be used in ways that honor the dignity of others and serve rather than manipulate, exploit, impersonate, or exclude them.

● Stewardship and Moderation: Data, time, attention, intellectual property, material resources, and creation are entrusted to our care. AI should be used prudently and proportionately, not compulsively or wastefully.

DELTA (University of Notre Dame) as a Complementary Discernment Tool

● Dignity: Does this AI use honor each person as made in the image of God, or reduce persons to data, outputs, rankings, or problems to be optimized?

● Embodiment: Does this AI use strengthen embodied learning, attention, and face-to-face relationships, or displace them?

● Love: Does this AI use foster authentic care and responsibility toward real persons? AI may simulate the language or appearance of care, but it cannot love.

● Transcendence: Does this AI use preserve openness to truth, beauty, goodness, wisdom, worship, wonder, and meaning beyond utility and optimization?

● Agency: Does this AI use strengthen the human capacity to attend, reason, listen, choose what is true and good, accept responsibility, and act with integrity, or does it surrender those tasks to a machine?

Cross-cutting Benedictine question

● Stewardship and the Common Good: Is this use moderate, equitable, protective of privacy and creation, respectful of human work and intellectual property, and ordered toward the good of the community, especially those who may be vulnerable or excluded?

A Four-Tier AI Policy Framework for Delbarton

The four-tier framework below identifies where AI may touch the learning process at Delbarton. The teacher determines the tier based on the assignment's learning objective or a particular component of it.

Students do not select their own tier. If a teacher or course policy has not designated an AI tier, students must ask before using AI on assessed work.

Two school-wide rules apply at every permitted tier:

1. Students remain responsible for the accuracy, integrity, appropriateness, and explainability of everything submitted.

2. Any permitted AI use must be disclosed according to the teacher’s directions. Disclosure never converts an otherwise prohibited use into a permitted one.

Tier 1, RED: Prohibited Uses

RED means that AI may not be used for the designated assignment, assessment, task, or component. The following uses are prohibited school-wide unless expressly authorized for a legitimate educational purpose:

● Submitting AI-generated content as one’s own work or concealing AI involvement when disclosure is required.

● Using AI on tests, quizzes, examinations, or other RED assessments.

● Using AI to perform the intellectual work that an assignment is specifically designed to assess, including writing, reasoning, translation, problem-solving, analysis, reflection, or coding when those skills are the learning objective.

● Inventing, altering, or fabricating sources, quotations, data, images, recordings, or evidence.

● Creating or distributing deceptive impersonations, deepfakes, synthetic voice or image content, or other media that misrepresent a real person without legitimate authorization.

● Using AI to harass, humiliate, sexualize, exploit, manipulate, discriminate against, or otherwise harm another person.

● Entering confidential, sensitive, personally identifiable, health, counseling, disciplinary, accommodation, personnel, safeguarding, or other protected information into an unapproved AI system.

● Using AI to evade school safety, security, filtering, monitoring, or academic-integrity controls.

● Using AI as a substitute for personal prayer, sacramental or pastoral care, spiritual direction, counseling, authentic friendship, or an assigned personal reflection whose purpose is the student’s own interior examination.

Academic integrity response

The School will address violations through its academic-integrity and disciplinary procedures. Responses should be proportionate to the seriousness, intent, and repetition of the violation and should include a formative component when appropriate; for example, a conversation with the teacher, demonstration of the student’s actual learning, revision or resubmission, or reflection on how integrity and trust can be restored.

Tier 2, ORANGE: Restricted Uses

ORANGE permits AI for pre-writing and inquiry support when the teacher has authorized it. The student must still perform the substantive research, reading, thinking, outlining, and drafting required by the assignment. In the STEM subjects, ORANGE permits the use of technology and AI to develop conceptual understanding. ORANGE does not permit the use of technology or AI for the generation of STEM problem set work and solutions, lab results or lab analysis in submitted work.

● AI may suggest search terms, research questions, concepts to investigate, possible counterarguments, or potential source leads.

● AI may be used for brainstorming or to test a student-created thesis, claim, or plan.

● AI is not a scholarly database and should not be treated as an authoritative source finder. Students must independently locate, read, evaluate, and verify all sources through appropriate primary texts, library databases, or other authoritative resources.

● No AI-generated prose, analysis, solution, translation, or other substantive content may appear in the final submission unless a different tier has been expressly designated for that component.

● Students must provide the required AI Use Statement or process log.

Tier 3, YELLOW: Restricted Revision Support

YELLOW includes all authorized ORANGE uses and permits AI to provide feedback on a student-created draft. The student, not the AI, must perform the revision. In the STEM disciplines, YELLOW permits the student to check answers and revise work based on errors found in the work.

● AI may flag grammatical errors, unclear sentences, organizational problems, repetition, possible gaps, or places where an argument is difficult to follow.

● AI may ask questions or identify areas for reconsideration; it may not supply replacement prose or new substantive content that the student inserts as his own.

● The student must compose all revised language himself and decide which suggestions to accept or reject.

● The student must retain and submit, when required, the pre-AI draft, version history, or other evidence of original authorship.

● STEM solutions and lab analysis/results may only be reviewed in AI when the teacher has explicitly permitted it by stating that the assignment is coded YELLOW

Tier 4, GREEN: Teacher-Directed Generative Use

GREEN permits AI-generated content only when the teacher explicitly designates the assignment or a component as GREEN because the educational objective includes learning to use, critique, direct, compare, transform, or evaluate AI-generated material.

● Students may use AI to generate designated portions of the work, up to and including a full work product when the assignment is expressly designed for that purpose.

● Students must critically evaluate AI output for accuracy, bias, omissions, fabricated information, ethical implications, and suitability for the task.

● Students must clearly identify AI-generated content and provide the required AI Use Statement, process log, prompt record, code comments, or other documentation specified by the teacher.

● Students remain responsible for every claim, citation, image, line of code, calculation, and other elements submitted and must be able to explain or defend the work without relying on the AI tool.

● GREEN should not be used when the primary learning objective is the very skill the AI would perform for the student.

Middle School AI Policy

As artificial intelligence continues to evolve, Delbarton is committed to ensuring that our students learn to use these tools thoughtfully, ethically, and responsibly. To keep the policy simple & easy for our middle school boys to understand, Delbarton will use a binary tier system, with assignments labeled RED or GREEN to indicate allowed AI use. The Middle School will not use the Orange or Yellow tiers. We will introduce these tiers to boys once they enter high school.

Below is an explanation of the Red & Green tiers as applied in the Middle School:

RED: AI Not Permitted
AI may not be used for the assignment. Students are expected to complete the work independently and demonstrate their own knowledge and skills.

GREEN: AI Permitted
AI may be used when the teacher allows it as part of an assignment. Students should evaluate AI’s output, recognize its strengths and limitations, and clearly identify when they used AI. Students are responsible for their final work and should be able to explain it without relying on AI.

Delbarton will continue to reinforce these expectations in the classroom and provide guidance to ensure students are using AI responsibly and in accordance with school policy.

Standard AI Use Statement

Unless a teacher specifies another format, a student who uses AI on an assessed assignment should include a brief disclosure such as:

AI Use: Flint was used to generate search terms and test my working thesis. I independently located and read the sources. No AI-generated text appears in this submission.

The AI Use Statement documents process; it does not replace normal citation of the actual sources, authors, data, or primary texts on which the work relies.

Four-Tier Summary

Tier	Permitted Role of AI	Student Obligation	Key Boundary
RED	No AI on the designated work.	Complete the work independently.	AI may not touch the assessed process or product.
ORANGE	Inquiry, research leads, brainstorming, testing ideas.	Verify sources; do the reading, thinking, outlining, and drafting.	No AI-generated substantive content in the final submission.
YELLOW	Feedback on student-created work.	Student performs every revision and preserves original draft evidence when required.	AI may diagnose or question; the student writes the revision.
GREEN	Teacher-directed generation, transformation, or task completion.	Disclose, verify, critique, explain, and take responsibility for all output.	Only when explicitly assigned; not when AI would replace the skill being assessed.

Applying the Framework

Example 1: RED

AI writes an assessed essay

A student enters an essay prompt into an AI tool, receives a five-page essay, and submits it for credit as his own work.

● Dignity: The student fails to honor and exercise the intellectual capacities entrusted to him and treats the teacher as an output processor rather than a mentor in learning.

● Embodiment: The use bypasses the reading, writing, discussion, feedback, and vulnerability that support learning in the student-teacher relationship.

● Love: The concealed use damages trust and substitutes simulated communication for the student’s own encounter with the teacher and subject.

● Transcendence: It bypasses opportunities for discovery, wonder, and meaning that often arise through the difficult process of forming an argument.

● Agency: It surrenders reasoning, authorship, and moral responsibility to a tool and constitutes academic dishonesty.

Summary: RED. AI has replaced the intellectual work and relational process that the assignment is designed to form.

Example 2: ORANGE

AI supports research planning

A junior writing a U.S. History paper asks Flint to suggest search terms and possible scholarly sources on the 1919 Red Scare and to identify weaknesses in his working thesis. He then independently locates and reads the sources through library databases, rejects irrelevant or unreliable suggestions, revises the thesis in his own words, outlines the paper himself, and drafts with no AI-generated prose. He submits the required AI Use Statement or process log.

● The student remains the author and exercises judgment rather than outsourcing it.

● AI accelerates discovery but does not replace reading, annotation, analysis, drafting, or teacher conference.

● Every source suggestion is independently verified; a model-generated citation is never accepted merely because it looks plausible.

Summary: ORANGE. This tier is explicitly for pre-writing investigation and pre-problem solving skill development. No AI-generated prose, problem solving, coding, or any other assessed work shall be part of the submitted work.

Example 3: YELLOW

AI flags revision needs

A senior writes a complete history essay (thesis, evidence, analysis, organization, and conclusion) without AI. He then asks Flint to flag unclear sentences, comma splices, repetition, and passages where the argument is difficult to follow. He decides which observations are useful and rewrites each sentence himself. He submits the pre-AI draft or version history if required.

● AI acts as a diagnostic reader, not a co-author.

● The student preserves his own voice by writing the revised language himself rather than installing AI-generated replacements.

● Peer review and teacher feedback remain preferred relational forms of revision; AI supplements rather than displaces them.

Summary: YELLOW. The governing line is: AI may help identify where revision is needed for clarity; the student must decide and write what the revision will be and make the revisions himself.

Example 4: GREEN

AI output becomes an object of study

Students research an ethical dilemma using authoritative sources. The teacher then directs them to use AI to generate a case study based on their research. Students compare the generated scenario with their sources, identify bias, omissions, and oversimplifications, revise it, and debate the resulting case in class.

● AI generation is explicit and teacher-directed.

● The intellectual objective is not merely to produce the case study but to evaluate, refine, and morally reason about AI-generated material.

● The activity returns to embodied discussion, shared inquiry, and accountable human judgment.

Summary: GREEN. AI-generated content is permitted because generation itself is part of the learning design and the students remain responsible for critical evaluation.

Faculty, Staff and Student Guidelines

AI Implementation Committee

Data Privacy and Stewardship

Human Review and Due Process

Approved AI Tools

Assessment, Accountability and Continuous Improvement

regular assessment

At least annually, the AI Implementation Committee should review the effect of AI policy and practice on student learning and formation, faculty experience, academic integrity, equity and accessibility, privacy and security, the teacher-student relationship, and alignment with Delbarton’s Foundation Documents and the DELTA framework.

Feedback and Recourse

● Student feedback, with anonymous options where appropriate.

● Faculty and staff discussion forums and discipline-specific review.

● Parent communication and opportunities for input.

● A clear channel for reporting concerns about bias, privacy, safety, academic integrity, or inappropriate AI-supported decisions.

● Meaningful human review and correction when an individual believes an AI-supported process has produced an inaccurate, unfair, or harmful result.

Incident Response

When an AI system causes or materially contributes to harm, the School should take proportionate immediate action, which may include suspending the AI tool’s use; investigating the cause and scope; protecting and remediating affected individuals; communicating appropriately and transparently consistent with privacy and legal obligations; and revising policy, training, or technical controls to reduce recurrence.

Continuous Improvement

This policy is a living document. Delbarton will review the core policy at least annually, and revise it when educational experience, Catholic teaching, legal requirements, vendor practices, or technological developments materially warrant change. Operational materials, such as approved tools, examples, and technical guidance, may be updated more frequently. Benedictine humility requires acknowledging when an approach is not working and having the courage to change course.

Conclusion: Formation for an AI Age

Delbarton is not simply managing technology. We are forming young men who must learn to live wisely and faithfully in a world increasingly shaped by artificial intelligence. Our goal is not to produce students who are merely competent with AI, but graduates who love learning, seek truth, work diligently, listen with the ear of the heart, exercise freedom responsibly, honor the dignity of others, practice moderation, care for creation, and serve the common good.

AI can assist learning when it remains a tool within the human relationships and purposes of education. It must not become a substitute for thinking, authorship, conscience, friendship, community, pastoral care, prayer, or the teacher’s personal responsibility for students. Efficiency is a legitimate good, but it is not the highest good; some forms of struggle, practice, patience, presence, and human work are themselves part of education and formation.

The Delbarton Benedictine–DELTA framework therefore rests on four enduring commitments:

● Mission: AI use must serve Delbarton’s Benedictine Catholic purpose and the formation described in the Profile of a Graduate.

● Transparency and Truth: We disclose permitted AI use, verify information, distinguish generated content from authoritative sources, and refuse deception.

● Human Responsibility: Consequential judgments about persons remain human; students and adults remain accountable for what they submit, communicate, or decide.

Discernment for the Common Good: We judge technology not simply by capability or efficiency but by whether it honors dignity, strengthens relationships and agency, protects privacy and creation, and contributes to human flourishing and lifelong service to God and neighbor.

Download a copy of Delbarton School AI Policy

Delbarton AI Policy Framework Appendix

Custom Class: menu-container

Custom Class: menu-utility-container

Custom Class: header-container

Custom Class: header-utility-container

Custom Class: nav-utility-header

In This Section