Calvary Preparatory Academy

AI Strategy Research Packet · v1.0 · April 2026

Recommendations & Roadmap

A staged plan: 90 days, 12 months, 5 years, 15 years

Calvary Preparatory Academy | AI in Education Research Packet

This document is a staged plan, designed for Calvary Prep’s actual constraints (online-native, California-credentialed, Christian, 50-state student body, executive director and board governance model) and the timeline the data suggests is realistic.

The plan is built around a simple architecture:

  • Track A — AI-assisted education. Use AI to deliver more personalized, more rigorous, more teacher-leveraging Christian education. Independent learning programs with human teacher oversight, modeled on the Brookings/CMU hybrid evidence base.
  • Track B — AI literacy education. Equip every Calvary Prep student with the technical capability and theological wisdom to use AI well in college and the workforce. A formal, grade-banded curriculum strand built on the OECD/EC framework plus Calvary Prep’s “Steward” axis.

Both tracks share infrastructure (policy, teacher PD, parent communication, evaluation rubric). Both ground in the W.I.S.E. Framework as the shared theological grid.

Phase 0 — The next 90 days (April–July 2026)

The goal of Phase 0 is to make all later phases possible. Almost no money is spent. Decisions are made.

1. Convene the Calvary Prep AI Working Group.

Owner: Jared Beck. By: May 15, 2026.

Composition (target ~7 people): - Jared Beck (chair, executive director) - Kim Beck (board member) - One academic leader (curriculum or principal-equivalent) - One classroom teacher (preferably one already using AI informally) - One technical advisor (Nomion AI, plus one independent voice if available) - One parent (chosen by Jared/Kim, ideally one who works in tech or education) - One student (an academically strong junior or senior — their voice is genuinely needed)

Operating rhythm: monthly meetings, working sessions between. Mandate: deliver to the full board by August 15, 2026 a v1 AI policy and a Phase 1 pilot plan, in time to communicate to families before the 2026–27 school year starts.

2. Adopt the W.I.S.E. Framework as the formal evaluative grid.

Owner: AI Working Group. By: First meeting.

Read iace.education before the first working session. Adopt with light adaptation (add the “Steward” axis from 03_christian_distinctive.md). Every AI tool the school adopts hereafter is evaluated against this grid, in writing, before deployment.

3. Read the foundational evidence base.

Owner: AI Working Group. By: Second meeting.

Required reading before substantive policy work: - Stanford SCALE — The Evidence Base on AI in K-12 - Brookings — What the Research Shows About Generative AI in Tutoring - ACSI/Cardus — Navigating AI in Christian Schools - The Gospel Coalition — Don’t Hand Education Over to AI (the loyal opposition voice — read it seriously) - Pew Research — How Teens Use and View AI - The summary documents in this packet (02_strategic_landscape.md, 03_christian_distinctive.md)

4. Survey current Calvary Prep AI reality.

Owner: Academic leader. By: End of June 2026.

You cannot build a thoughtful policy without knowing what’s already happening: - Which teachers are already using AI? For what? - Which students are using AI? For what? How transparent are they being? - What tools are already paid for or implicitly approved (Gemini, Khanmigo, ChatGPT Edu, etc.)? - What does our current academic-integrity language say, and how does that map to current student practice?

This survey is non-punitive. Frame it explicitly: “We’re building a Christian-distinctive AI strategy for Calvary Prep, and we need to start from honest information.” The Stanford SCALE finding that 50%+ of teacher AI prompts are content-generation tasks is what this survey will likely confirm internally.

5. Pick the Phase 1 pilot subject.

Owner: Working Group. Recommendation: High school math.

Math is the right pilot subject because: - The strongest tutoring evidence (Brookings, Khanmigo, LearnLM, Nature RCT, CMU year-long study) is in math - The harm signal from unstructured AI (Wharton 17% drop) is also in math, so the contrast between “good” and “bad” AI tutoring is sharpest and most teachable - It is one of the easiest subjects in which to design assessments that are AI-resilient (oral defense, in-class problem solving, process portfolios) - Calvary Prep’s online infrastructure already mediates math instruction

Secondary candidate: Bible/theology, where the W.I.S.E. axes can be most clearly demonstrated.

Phase 1 — Foundation (August 2026 – June 2027)

The goal of Phase 1 is to ship one pilot well, build teacher capacity, and earn parent trust. Underpromise. Overcommunicate.

1. Launch the math AI tutoring pilot.

Design principles (drawn from the Brookings/CMU hybrid model): - One scaffolded AI tutor (recommended for evaluation: Khanmigo as the most extensively studied and most theologically inert option for math; MagicSchool.ai for teacher-facing work; or a custom Claude-based tutor configured by Nomion AI) - Required human teacher supervision in every interaction loop. AI suggests; the teacher reviews; the student receives. Aim for the LearnLM 76% approval rate as a quality benchmark. - Process portfolio assessment — students show their work over time, not just their answers. Periodic oral defenses over video. - Explicit AI-use declarations. Students learn to say “here is how I used AI on this problem and here is what I did myself.” Transparency is taught, not policed. - Pre/post measurement of learning outcomes against a non-pilot comparison cohort.

Scope: One grade level, one math course, one or two teachers. ~30–60 students.

2. Build out teacher AI professional development.

Owner: Academic leader + technical advisor.

The single binding constraint on AI integration is teacher capacity. The SREB four-phase model (build, explore, apply, iterate) is the right scaffold (sreb.org).

For Calvary Prep specifically: - All teachers complete a 4-session AI literacy + theology of technology course (including the W.I.S.E. Framework) - All teachers receive supervised practice with the school-approved AI tools (likely MagicSchool.ai for lesson prep, Gradescope or similar for grading support) - A small group of “AI lead teachers” go deeper and serve as peer support - Time savings (the Gallup data suggests ~6 hrs/week for weekly users) are explicitly reinvested in individualized student feedback and discipleship conversations — not banked. Make this commitment public to families.

3. Begin Track B: AI literacy curriculum design.

Owner: Curriculum lead + working group. Goal: v1 grade-banded sequence ready for fall 2027.

Use the OECD/EC framework’s four dimensions (understand, use, evaluate, create) plus Calvary Prep’s Steward axis. Map content into existing courses rather than building a single elective: - Understand — covered in computer/tech course units, with a Genesis 1–2 anthropology module - Use — embedded in writing, research, and project work across subjects - Evaluate — primary home is in Bible/theology and rhetoric/logic classes (bias detection, source evaluation, theological vetting) - Create — senior capstone project: every Calvary Prep graduate ships one AI-built artifact (a study tool, a ministry resource, a research project) with documentation - Steward — woven through Bible classes, chapel, and parent communications; assessed in capstone reflection

ETS’s K-12 framework (rr.ets.org) provides grade-banded skill progressions you can adapt.

4. Redesign academic integrity policy.

Owner: Academic leader.

Replace detection-based language with process-based language: - Students are expected to declare AI use in writing on every major assignment - Every major assignment includes a process artifact (drafts, oral defense, in-class component, screen recording, or annotation reflection — pick what fits the task) - AI use without declaration is the integrity violation, not AI use itself - The W.I.S.E. axes are taught as the framework students use to reflect on their own AI use

This is more rigorous than detection-based policies, not less. It assumes students will use AI and asks them to use it as image-bearers. It also matches where the higher-education world is moving.

5. Write the parent-facing AI letter.

Owner: Jared Beck, with working group input. Send: Before the 2026–27 school year begins.

Three things this letter does: - Explains Calvary Prep’s posture (using the paragraph in 03_christian_distinctive.md) - Explains what is changing in 2026–27 (math pilot, integrity policy update, teacher PD investment) - Invites parent engagement and questions

The PDK/EdWeek finding that public support for AI in lesson planning dropped 13 points in one year (edweek.org) tells you exactly how important this letter is. Don’t let parents discover Calvary Prep’s AI strategy through their kids. Tell them first.

Phase 2 — Integration (2027–2028)

The goal of Phase 2 is to scale what worked in Phase 1 and ship the AI literacy curriculum.

1. Expand AI tutoring beyond math.

Based on Phase 1 learning outcomes data, extend hybrid AI tutoring into 1–2 additional subjects (likely candidates: foreign language, science).

2. Launch v1 of the AI Literacy Curriculum across grade levels.

Every Calvary Prep student now receives AI literacy formation across the OECD framework + Steward axis. Senior capstone “ship one AI-built artifact” project becomes a graduation requirement.

3. Publish Calvary Prep’s AI policy and outcomes data.

This is the moment Calvary Prep becomes visible as an AI-fluent Christian school. Publish: - The W.I.S.E.-based AI policy (with Steward axis added) - Anonymized Phase 1 outcomes data - Sample student capstone projects - Theological frame as a usable document for other Christian schools

This is also a strategic admissions and donor moment. There is currently no online Christian high school in the country with a public, evidence-based AI program. Calvary Prep can be the first to claim that position.

4. Build relationships with Christian AI institutions.

  • Send a faculty representative to the Baylor Symposium on Faith and Culture
  • Reach out to Tim Hwang’s Institute for Christian Machine Intelligence for advisory dialogue
  • Engage with ACSI on their evolving AI work
  • Consider hosting or co-hosting a Calvary Prep / Nomion AI symposium on AI and Christian secondary education

5. Begin tracking graduate outcomes against AI-fluency markers.

What share of Calvary Prep graduates enroll in college programs that explicitly value AI capability? What share secure AI-related internships? What share say in exit surveys that they feel prepared for the AI-saturated workforce? This becomes admissions-narrative gold over time.

Phase 3 — Leadership (2028–2031)

The goal of Phase 3 is to make Calvary Prep one of the schools setting the standard for Christian secondary education in the AI era.

1. Launch a Calvary Prep AI Track or Distinction.

Beyond the universal AI literacy curriculum, students who go deeper graduate with a formal AI Distinction. Curriculum elements: structured exposure to RAG, agent frameworks, vector databases, model evaluation, prompt engineering, and AI-assisted product building (the technical cluster employers actually want — see 02_strategic_landscape.md §5). All taught within the W.I.S.E. theological frame.

2. Develop an AI-Christian-Mentorship Program.

Connect Calvary Prep students with Christian working professionals in AI: Anthropic researchers, Microsoft Christian community, Christian AI ethics academics, etc. There is real demand; there is essentially no supply pipeline. Calvary Prep can build it.

3. Productize the curriculum.

The AI literacy curriculum Calvary Prep builds for itself can be licensed to other Christian schools. ACSI member schools (~24,000 globally) are looking for exactly this. This is a sustainability and reach play, not just a revenue play.

4. Publish research.

Calvary Prep can be the first online Christian K-12 school to publish longitudinal data on Christian formation outcomes in an AI-integrated environment. The MDPI Religions journal, the Journal of Christian Education, and the Cardus Education Survey are all natural homes. The academic literature on this is currently almost entirely prescriptive — actual empirical data from a real school would be a significant contribution.

Phase 4 — Horizon (2031–2041)

The 15-year horizon is the one Jared and Kim explicitly raised. We will not pretend to predict it precisely. The technology in 2041 will be substantially different from anything available in 2026. But the directional commitments that should hold:

1. AI capability becomes a baseline expectation, not a distinction.

Just as computer literacy became a baseline by 2010, AI fluency will be a baseline by 2030. Calvary Prep’s distinction will need to keep moving up the value chain — from “we teach AI literacy” to “we teach AI stewardship” to whatever the 2035 frontier of Christian AI formation looks like. The W.I.S.E. axes are durable; the specific technologies are not.

2. The Christian distinctive becomes more, not less, valuable.

The Anthropic-consults-Christian-leaders moment of 2026 is the early signal. As AI systems become more capable, society will need wisdom traditions that can think about agency, consciousness, dignity, and stewardship at depth. Christian theology is one of the strongest such traditions on offer. Schools that form students fluent in both AI and Christian theology produce graduates whose voice will be increasingly sought.

3. Independent learning programs become the norm in Christian education.

The economic pressure on Christian schools (declining enrollment, teacher shortages, tuition affordability) combined with the Brookings/CMU hybrid evidence base means most Christian education in 2041 will involve some form of AI-assisted independent learning under human teacher oversight. The schools that learn to do this well in 2026–2028 will be the ones the rest of the field looks to in 2030–2035.

4. Calvary Prep’s online-native model becomes the dominant model.

Brick-and-mortar Christian education will continue, especially in concentrated population centers. But the addressable market for high-quality online Christian secondary education is structurally larger than for any single brick-and-mortar institution, and it grows with every Christian family that prioritizes formation but cannot relocate. Calvary Prep’s 2009 thesis is more correct now than when it was launched. The 2041 question will be whether Calvary Prep stayed at the front of that model or got passed by schools that started later but moved faster.

What Calvary Prep should not do

Brevity matters here. Three things to avoid:

1. Don’t adopt AI access as a policy.

“We allow students to use AI” is not a policy. It is the absence of one, and it will produce the Wharton 17% problem in your own students. Specific tools, specific design, specific supervision.

2. Don’t lead with technology adoption announcements.

Most schools that have made public AI announcements have announced tools. Calvary Prep should announce posture (the Christian distinctive) and outcomes (the pilot data, the capstone projects, the formed graduates), with tools as the implementation detail. The story is the school’s mission, not its software.

3. Don’t try to do this without budget.

Phase 1 costs real money — primarily teacher PD time, working group time, and tool licensing. The order-of-magnitude estimate for Phase 1 (Aug 2026–June 2027) is somewhere between $30k and $80k depending on how much teacher PD is built internally vs. brought in. Phase 2 scales with student-count. This is small relative to a year’s tuition revenue and large relative to what most schools are spending on this. It is the highest-leverage formation investment available to Calvary Prep right now.

Decision points the board needs to make

In rough order of urgency:

  1. Approve standing up the AI Working Group and naming Jared as chair. (May 2026)
  2. Approve adopting the W.I.S.E. Framework + Steward axis as Calvary Prep’s evaluative grid. (June 2026)
  3. Approve a Phase 1 budget for the math pilot, teacher PD, and curriculum development work. (June–July 2026)
  4. Approve and review the parent-facing AI letter before it goes out. (July 2026)
  5. Approve the v1 AI policy for the 2026–27 school year. (August 2026)
  6. Review Phase 1 outcomes and approve Phase 2 scope. (June 2027)

If the board approves only #1 in the next 30 days, the rest can follow on a working timeline. If #1 doesn’t happen by mid-May, the 2026–27 school year will start without a coherent Calvary Prep AI posture, and the gap will widen.

“Where there is no revelation, the people cast off restraint; but happy is he who keeps the law.” — Proverbs 29:18 (NKJV)

The schools — and graduates — who will flourish in the AI era are those who remember what restraint is for. That is the formation Calvary Prep already exists to deliver. AI is the new context in which it must be delivered. The work is the same; the tools have changed.