Risks & Objections · CPA AI Strategy

Calvary Preparatory Academy | AI in Education Research Packet

A serious AI strategy has to engage the serious objections rather than route around them. Several of the objections below are theologically substantial; some are pedagogical; some are practical. We give each a fair hearing and indicate how the proposed Calvary Prep approach responds to it.

1. The pneumatological objection: AI shortcuts the work of the Holy Spirit.

The argument. Genuine Christian study, prayer, and Scripture engagement involve the work of the Holy Spirit through human struggle, dependence, and patience. AI accelerates the output of those activities (a written reflection, a sermon outline, a Bible study summary) without producing any of the interior fruit. By embedding AI in Christian education, schools risk training students to bypass exactly the kind of slow, prayerful struggle that genuine spiritual formation requires.

This is articulated most sharply by grassroots Christian voices on social media (X/@RunyaMahtan; X/@thetangles) and gestures toward a longer tradition of Christian thought (Bonhoeffer, Lewis on the dangers of technological substitution for inner work) that Calvary Prep should not dismiss.

The serious response. The objection is correct about the danger but wrong about the conclusion. The actual danger is substitution — using AI in place of the inner work. Refusing AI does not solve substitution; it just removes one specific instance of it. The same substitution can happen with commentaries, sermons, podcasts, study Bibles, or any other resource that mediates between the student and the text.

The right response is to architecturally enforce the distinction between AI and inner work:

AI is used for information tasks: summarizing background, generating practice problems, drafting initial outlines, surfacing relevant cross-references.
Inner work — meditation, prayer, memorization, the slow chewing-over of a passage, the wrestling with a hard truth — is protected from AI by design. The student does not get to outsource it. Assessments are structured so they can’t.

This approach explicitly serves the pneumatological concern rather than ignoring it. Done well, it produces students who are more spiritually formed than they would be in an “AI-free” school that hasn’t thought carefully about the same dynamic in its commentary use, podcast diet, or YouTube sermon habits.

The W.I.S.E. axes — particularly Worship — name this directly. Whatever is being substituted for genuine dependence on God is the actual problem, AI or not.

2. The discipleship-formation objection: AI cannot impart wisdom or virtue.

The argument. This is the Gospel Coalition position (thegospelcoalition.org) and it is the strongest institutional voice of caution in the field. Education at its best is not the transmission of information but the formation of a person — wisdom, virtue, character, a way of seeing the world. Teachers do that through relationship, modeling, and patient correction over years. AI cannot. So embedding AI in schools risks shifting the school’s center of gravity from formation to information without anyone noticing.

The serious response. Calvary Prep agrees with the underlying claim. Wisdom is not transmitted by AI. The argument is not against AI; it is against schools letting AI displace the things AI cannot do. The proposed strategy holds the formation work as the school’s center and uses AI to free teachers for it.

The CMU year-long study of human + AI tutoring found that the additional 0.36 grade-level gain came from human supervision, not from the AI alone. The Gallup/Walton finding that 80% of AI-using teachers report improved work quality is paired with the finding that the time saved is being reinvested in individualized student feedback. Formation work doesn’t shrink in an AI-augmented school — it can actually grow, if the architecture protects it.

The risk is real if a school adopts AI without thinking. Calvary Prep is not doing that.

3. The cognitive-development objection: AI degrades student thinking.

The argument. This is the strongest empirically-supported objection. The Wharton/PNAS finding that students with open ChatGPT access scored 17% worse on independent exams (X/@asanwal) and the MIT/Oxford/CMU finding that AI assistance reduces persistence (X/@rohanpaul_ai) are not opinions — they are randomized controlled trials. Christian schools should not subject their students to interventions that have been shown to harm independent reasoning.

The serious response. This objection is correct as a critique of unstructured AI access. It is incorrect as a critique of scaffolded, supervised AI tutoring. The same body of research that produces the 17%-worse finding also produces the LearnLM trial, the CMU year-long study, and the Nature Scientific Reports RCT — all of which show learning gains under specific design conditions.

Ethan Mollick’s framing is the most useful synthesis: unconstrained AI shortcuts learning; tutor-prompted AI improves it. (X/@emollick)

Calvary Prep’s strategy should explicitly refuse the unstructured-access pattern that produces the harm. The pilot is built on the design conditions associated with gain (scaffolded tutor, human supervision, process portfolios, transparent use). The math pilot is specifically the most-studied subject for this reason — there is enough RCT data to design a pilot that is unlikely to harm and reasonably likely to help.

4. The parental-trust objection: parents will think we’ve gone soft on rigor.

The argument. Calvary Prep parents pay tuition because they want their kids to read hard books, write hard papers, do hard math, and be formed by serious teachers. If they hear “we are using AI in math class,” many will hear “we are letting students take shortcuts.” The PDK/EdWeek poll finding that public support for AI in lesson planning dropped from 62% to 49% in a single year, and that 70% of parents oppose AI tools accessing student data (edweek.org), is a real warning.

The serious response. Parents are right to be skeptical of schools that adopt AI sloppily. The proposed Calvary Prep approach should be communicated to parents first and presented as the opposite of softening:

AI does not replace the rigor; it increases it. Process portfolios, oral defenses, and AI-use declarations are more demanding of students than a take-home essay. They cannot be faked.
Teachers are reinvesting saved time into more individualized feedback per student than they could deliver before. Each student gets more teacher attention, not less.
The Christian formation frame is foregrounded. Parents see exactly what theological commitments shape Calvary Prep’s AI use, and what their kids are being formed in.

The parent letter (Phase 1, recommendation #5 in 04_recommendations_roadmap.md) is the central trust-building artifact. If it is honest, specific, and theological, it builds trust. If it is buzzword-heavy or vague, it erodes trust. Get this letter right.

5. The academic-integrity objection: AI makes cheating undetectable.

The argument. Turnitin reports 15% of submitted essays are now more than 80% AI-generated, up from 3% in 2023. The University of Arizona, Cornell, and Pitt have abandoned AI detection because it produces unacceptable false-positive rates against ESL and neurodiverse students. (KOLD) Calvary Prep cannot detect cheating reliably and so cannot maintain academic standards.

The serious response. The objection assumes detection is the only path to integrity. It isn’t, and it is increasingly the wrong path.

Calvary Prep’s response is to redesign assessment such that AI use is either explicit (and therefore not cheating) or impossible to substitute for the assessed work. Online schools have a structural advantage in this redesign: oral defenses over video, process portfolios in the LMS, in-class timed work using monitored sessions, and transparent AI-use declarations are all easier to architect when assessment is already digital.

The expert consensus has tipped this direction. Both the Washington Post and Inside Higher Ed published opinion pieces on the same day (April 13, 2026) calling detection-first integrity strategies harmful and unsustainable. Calvary Prep should not invest in the dying paradigm. It should build the new one early.

6. The “AI will be different in five years” objection.

The argument. Whatever Calvary Prep builds in 2026 will be obsolete by 2031. The technology is moving too fast. Wait until things stabilize.

The serious response. The technology is moving fast. The theological architecture (W.I.S.E. + Steward) is durable across technology generations. The pedagogical pattern (scaffolded, supervised, process-based) is durable across tools. The cultural moment (Christian schools defining their posture) will not wait — it is being defined right now, and the schools that wait will inherit defaults set by schools that didn’t.

Calvary Prep’s roadmap is deliberately built around durable principles with tool-agnostic implementation. When the math AI tutor changes in 2028, the W.I.S.E. evaluation grid still applies. When prompt engineering becomes obsolete in 2029, the AI literacy curriculum’s Understand/Use/Evaluate/Create/Steward skeleton survives the change. This is what makes the strategy a strategy and not a tool-adoption plan.

7. The “we are not a tech school” objection.

The argument. Calvary Prep is a Christian school first. Why should we organize anything around AI? We don’t organize around any other technology.

The serious response. AI is a different category of technology than the ones the objection is implicitly comparing it to. It is closer to literacy than to email. It changes how human beings think, work, and learn — which is exactly the territory a school operates in. The OECD/EC framework, the IMF, and the Brookings Institution are all treating AI as an infrastructure-level shift in education, not a feature.

Also, the framing inverts: Calvary Prep is precisely not organizing around AI. It is organizing around Christian formation in an era when AI is a defining feature of the world that formation has to engage. The school’s center is unchanged. The context isn’t.

8. The cost objection.

The argument. Phase 1 alone is $30k–$80k. Phase 2 scales with enrollment. Calvary Prep is a small school with budget pressures like every Christian school. Where does this money come from?

The serious response. This is a real question and deserves a real answer in financial planning. Three notes:

Phase 1 is mostly teacher time, not tooling. The largest line item is teacher PD investment. That’s a one-time front-loaded cost; Phase 2 onward, PD is amortized over a much larger ongoing benefit.
The alternative is also expensive. A school that does nothing in 2026–27 and then has to scramble in 2028 to build a coherent AI strategy under competitive pressure will spend more, less effectively, with worse parent trust.
There may be funder appetite specifically for this work. ACSI grants, Murdock Trust, Lilly Endowment, and faith-aligned tech philanthropies are all funding Christian education / AI work in 2025–26. Calvary Prep’s online-native model + Christian distinctive + clear theological architecture is exactly the kind of project that lands well with these funders. Worth a serious development effort.

9. The “students are already doing this on their own” objection.

The argument. 84% of high school students are already using AI (Pew). Whatever Calvary Prep does or doesn’t do, students are figuring it out themselves. Why does the school need to do anything formal?

The serious response. Yes — they are figuring it out themselves, badly, in exactly the unstructured-access pattern the Wharton study identified as harmful. The 64% of teens who report receiving no AI education at school is a description of the problem, not a solution. (Junior Achievement) The school’s job is to put a Christian formation frame around the activity students are already doing. If we don’t, the algorithm and the peer group will.

10. The reputational objection: AI is a culture-war flashpoint.

The argument. Some Christian parents have strong negative associations with AI (job loss, deepfakes, manipulation, transhumanism, prophecy concerns). Going public with an AI strategy could lose enrollments.

The serious response. This is a real concern and the parent communication strategy needs to engage it carefully. Two things to know:

The parent population is heterogeneous. Some Calvary Prep families will be technology-skeptical; others will be technology-enthusiastic; most are somewhere in between. A clear Christian frame that names the concerns and answers them well is more defensible than vagueness.
Schools that don’t go public will end up communicating their posture by default through how their students show up. A graduate who can’t compete with AI-fluent peers in college will say more to prospective families than any letter.

The right reputational frame is Christian seriousness in the AI era, not technology adoption. That frame holds up under scrutiny from any direction on the parent spectrum.

What we are not going to do, even if asked

Three commitments worth stating explicitly:

No AI evaluation of student spiritual development. AI does not assess discipleship. Period. That work belongs to teachers, mentors, parents, and the student’s own conscience before God.
No black-box adoption of any AI tool that claims to do “biblical AI.” Every tool, including faith-aligned tools like Truthly or Halo, is evaluated against the W.I.S.E. + Steward axes before deployment. “Christian-branded” is not a substitute for theological evaluation.
No use of AI to surveil students. Online schools have particular pressure to deploy invasive monitoring. Calvary Prep should explicitly reject that path. Trust is built through formation and transparent assessment design, not through surveillance.

“Wisdom is the principal thing; therefore get wisdom. And in all your getting, get understanding.” — Proverbs 4:7 (NKJV)

The objections above are wisdom from various directions. Engaging them seriously is part of the work, not an obstacle to it.