Calvary Preparatory Academy

AI Strategy Research Packet · v1.0 · April 2026

Executive Summary

The five-minute version for Jared and Kim

Calvary Preparatory Academy | AI in Education Research Packet

The situation

K–12 education has crossed a structural threshold in the last 18 months. The data points are no longer projections; they are baselines:

  • 84% of U.S. high school students already use generative AI for schoolwork. Pew Research, February 2026 (n=1,458 teens). Among teens specifically, 54% use AI chatbots for schoolwork. (Pew)
  • 60% of U.S. K–12 teachers used AI in the 2024–25 school year. 30% are weekly users; weekly users save ~6 hours per week — roughly six full work weeks per year. (Gallup/Walton)
  • Demand for AI skills in entry-level job postings nearly tripled in six months. Now over 33% of entry-level postings explicitly require AI proficiency, up from ~11% in fall 2025. AI-skilled workers command a 56% wage premium across sectors. (NACE, April 2026; PwC Global AI Jobs Barometer)
  • Only 5.4% of students receive any formal AI literacy training, despite 80% of schools claiming to teach AI basics. (ResearchGate integrative review; EdWeek)
  • Boston became the first major U.S. city to mandate AI literacy for all public high school students in March 2026. State and federal mandates are likely to follow within 24 months.

The fork

Two things are simultaneously true and create the central tension every school is navigating:

1. Pedagogically designed AI helps students. A Nature Scientific Reports RCT (n=316) found AI tutoring produced double the learning gains of in-class active learning, in less time. Google’s LearnLM trial in UK secondary schools beat human-only tutoring by 5.5 percentage points on novel-problem transfer. A year-long CMU study of 350+ seventh graders found human-supervised AI tutoring delivered an additional 0.36 grade-level gain over AI-only deployment.

2. Unstructured AI access actively harms students. A Wharton/PNAS study found high school math students given open ChatGPT access scored 17% worse on subsequent exams than peers who used no technology — despite showing 45% gains during AI-assisted practice. Students didn’t know it was happening. An MIT/Oxford/CMU study found AI assistance reduces persistence on hard problems and degrades independent problem-solving over time.

The question for every school in the country has shifted from “should we use AI?” to “how do we use AI without quietly hollowing out the formation we promised parents we’d deliver?”

What’s at stake for Calvary Prep specifically

You were built around a structural advantage in 2009 — a serious online Christian education delivered to families across the country — that COVID then made the dominant model. That advantage has been compressed. Sevenstar, Liberty University Online Academy, FLVS, Whitmore, and dozens of full-time online private schools now compete in your space. Most of them are publicly Christian or Christian-friendly. The thing that distinguishes Calvary Prep from the broader online K–12 market is not the medium anymore — it is the formation.

That formation thesis is now under unique pressure. Three forces converge specifically on online Christian schools:

  1. Workforce repricing. Calvary Prep’s parents are paying tuition because they expect their kids to enter college and the workforce competitive with peers from the strongest secular schools. The job market is repricing around AI fluency now — not in 2030. A 2026 graduate without demonstrable AI capability is entering the worst entry-level market in five years (42.5% underemployment, the highest since 2020). A 2026 graduate with it is entering a talent-scarce premium market.

  2. Academic integrity collapse in detection-based models. The University of Arizona, Cornell, and Pitt have all disabled or severely restricted their AI detection systems due to false-positive rates that are punishing honest students — including ESL students, students with formal writing styles, and homeschoolers. Detection is dead. The schools that adapt fastest will be those that redesign assessment around process, oral examination, and portfolio rather than chasing detection. Online schools have a structural advantage in this redesign because their assessment infrastructure is already digital and re-architectable.

  3. Discipleship formation under quiet algorithmic pressure. The most serious theological objection to AI in Christian education — articulated by The Gospel Coalition’s Don’t Hand Education Over to AI and a sharper pneumatological critique from grassroots Christian voices — is that AI can transmit knowledge but cannot impart wisdom, virtue, or the relational texture of genuine discipleship. This objection deserves serious engagement, not dismissal. It is also the wedge that makes a Christian school’s AI policy fundamentally different from a secular school’s.

The opportunity

There is a 12–18 month window in which a strategically minded online Christian high school can occupy a position no one else holds: an evidence-based, theologically coherent, AI-fluent Christian high school. The W.I.S.E. Framework (IACE, February 2026) and the ACSI/Cardus survey (cardus.ca) provide foundational documents that did not exist 18 months ago. Tim Hwang’s Institute for Christian Machine Intelligence is producing peer-reviewed work arguing that Christian theology offers concrete technical contributions to AI alignment — meaning the conversation is no longer one in which Christian schools are reactive defenders. They can be missional shapers.

Specifically, three things are available to Calvary Prep that are not available to most schools:

  1. You are online-native. You can deploy AI tutoring infrastructure without retrofitting a brick-and-mortar institution. Your teachers already operate in digital workflows. Your students are already comfortable on screens. Most of the painful organizational change required for AI integration is change you have already absorbed.

  2. You serve students nationally. You can offer an AI-literacy curriculum that scales without geographic constraint. If Calvary Prep builds the strongest Christian AI literacy program in the country, the addressable market is every Christian family with internet access — not just families within driving distance of a campus.

  3. You have a Christian distinctive that is becoming more, not less, valuable. Anthropic, the maker of Claude, recently consulted ~15 Christian leaders on AI moral development. The 2026 Baylor Symposium on Faith and Culture has dedicated sessions on AI reimagining K–12 discipleship. Christian theology is being treated as a technical resource for AI alignment by major industry voices. The cultural moment for Christian schools to lead in AI ethics is genuinely now.

What we are recommending at the highest level

Three commitments to make in the next 90 days:

  1. Stand up a Calvary Prep AI Working Group (Jared, Kim, one teacher, one parent, one student, one technical advisor). Mandate: deliver a v1 AI policy and pilot plan to the board by August 2026, in time for the 2026–27 school year. Use the W.I.S.E. Framework as your starting document.

  2. Pilot one AI-assisted independent learning track in a single subject (recommend: high school math) for the 2026–27 school year, modeled on the Brookings hybrid-AI-tutoring evidence base. Human teacher oversight is not optional in this design — it is the source of the 0.36 grade-level gain.

  3. Begin building an AI literacy curriculum module for the 2026–27 school year, integrated across grade levels rather than offered as a single elective. Use the OECD/EC AI Literacy Framework’s four dimensions (understand, use, evaluate, create) plus an explicit fifth dimension Calvary Prep adds: steward — what does it mean to use AI as an image-bearer of God, with wisdom, restraint, and love of neighbor?

If Calvary Prep does these three things in 2026, the school will be among the first 5% of Christian high schools in the country with a coherent AI program — and will have a story to tell prospective families that no one else can tell yet.

What this packet contains

  • 02_strategic_landscape.md — the full evidence picture across tutoring, literacy, teacher tools, academic integrity, and workforce
  • 03_christian_distinctive.md — the theological architecture for a serious Christian AI program
  • 04_recommendations_roadmap.md — staged plan with concrete next steps
  • 05_risks_and_objections.md — honest engagement with the serious objections
  • 06_research_appendix.md — full source research and curated reading list

The supporting research from all six streams is preserved at Workshop/research/topics/ and linked throughout.

“The simple believes every word, but the prudent considers well his steps.” — Proverbs 14:15 (NKJV)

The data is unusually clear right now. The window for prudent action is open and not infinite.