AI in Higher Education — May 01, 2026 | The Langenkamp Dictionary

May 01, 2026 · Vol. 15

A weekly brief for the Management Department, Isenberg School of Management, UMass Amherst. By Matthew D. Langenkamp / 雷邁德, with research assistance from Thea 🪻✨.

← Back to archive · ← Back to Dictionary

Overview

A consequential week, with a clear theme: the bill is coming due on AI in higher education — for institutional governance, for assessment design, and for the value proposition of a degree. Arizona State drew faculty fury after quietly launching an AI tool that repackages their lectures for $5/month. EAB published student-survey data showing 4 in 10 college students say AI will shape their career choice — and 1 in 10 have already changed their major. AACSB’s 2026 Global Standards became official and take effect in 60 days. OpenAI shipped GPT-5.5 to ChatGPT Edu, raising the floor under what students can produce. And the New York Times documented a real pedagogical reversal — writing is moving back into the classroom because faculty can no longer trust take-home work. Below: seven items, one case study, and a new closing section connecting the week’s news to MGMT 494BI Business Policy & Strategy.

This issue introduces a recurring closer: From the Capstone — a brief connection from the week’s developments to one classroom question we are wrestling with in the strategy capstone.

1. Tools: GPT-5.5 Lands in University Edu Plans

OpenAI released GPT-5.5 and GPT-5.5 Pro on April 23–24, with deployment to ChatGPT Edu — the university-managed tier — effective immediately. The headline capability gains: a 1M-token context window (the equivalent of an entire textbook plus case-pack in a single prompt), substantially stronger reasoning, and improved tool use. For institutions on existing ChatGPT Edu contracts, the upgrade is automatic at no additional cost. (OpenAI, April 23, 2026; CNBC, April 23, 2026)

The practical implication for course design is direct: the GPT-5.5 student is not the GPT-4 student. Assignments calibrated to the prior generation — case write-ups, strategy memos, take-home analyses — may now be under-specified. A task that produced a recognizable distribution of student work two semesters ago may produce a flatter, more competent, more uniformly AI-assisted distribution this fall. The relevant question for assessment design is not whether students will use it (they will), but what learning we are assessing once they do.

For colleagues experimenting with AI integration: GPT-5.5’s longer context window changes what is feasible — full case-method instruction with AI co-analysis becomes practical in a way it wasn’t before. That is an opportunity, but it also raises the bar on what counts as student-generated insight.

2. Policy & Governance: AACSB 2026 Standards Are Now Final — Effective July 1

The AACSB 2026 Global Standards for Business Accreditation (and the companion 2026 Accounting Standards) were ratified on April 14 at the AACSB International Conference and Annual Meeting (ICAM) in Seattle, and the final standards documents are now published. Both take effect July 1, 2026 — sixty days from now. (AACSB.edu, April 2026; aacsb.edu/educators/global-standards)

Two elements are worth flagging for our department:

Standard 4.1 clarifies the expectation that curriculum “reflects current and emerging business theories, technologies, and practices.” This language gives AACSB peer-review teams an explicit hook to probe AI integration in curriculum design. It is not a directive to adopt AI — but it is a requirement to demonstrate that curriculum has thought about it.
Standard 5 (Faculty Sufficiency and Qualifications) sits behind the larger conversation about who is responsible for AI-mediated content in courses — see the ASU case study below.

Schools mid-cycle in their accreditation review can elect to complete under the 2020 standards, but new cycles begin under 2026. For Isenberg, the practical implication is that any self-study material we are producing now should be drafted with the new language in mind.

(Background on the broader 2026 changes — including the recalibration of DEI-related language and the consolidation of societal impact under Standard 9 — is in Appendix A.)

3. The Demand Side: 4 in 10 Students Say AI Will Shape Their Career — 1 in 10 Already Changed Major

EAB published findings on April 30 from a survey of 9,516 college-eligible students conducted February–March 2026. Headline results: 42% report that AI will influence which career they pursue, and 10% say they have already changed their planned major because of AI concerns. The dominant emotional registers in the open-response data: “uncertain,” “concerned,” “nervous,” “depressed.” EAB’s head of research framed the takeaway sharply: “AI is upending the value equation in higher education… colleges must prove they’re preparing graduates.” (Inside Higher Ed, April 30, 2026)

This compounds the Lumina-Gallup 2026 study released earlier in April, which found 47% of currently-enrolled students have seriously considered changing majors due to AI-driven labor-market fears, and 57% are using AI at least weekly despite institutional discouragement.

For our department: This is a demand-side signal with direct enrollment and advising implications. Students arriving in the fall will already be running their own internal cost-benefit analyses on whether the management major they applied to still makes sense in an AI economy. The advising conversations are about to change. So is the answer to “what does an Isenberg degree get me?” — at least for the cohorts thinking about it most seriously.

The question to our committees and curriculum leads is no longer “what is our AI policy?” — it is, as EAB puts it, “can we show our degree is worth it in an AI economy?” That is a different framing, and a more demanding one.

4. Assessment Reversal: Writing Moves Back Into the Classroom

The New York Times ran a major piece on April 30 documenting what is becoming a quiet pedagogical reversal: high school and college instructors, no longer able to trust that take-home written work reflects student understanding, are moving writing back into the classroom. In-class essays, oral examinations, and handwritten work are being rehabilitated as forms of authentic assessment for the first time in a generation. (New York Times, April 30, 2026 — paywalled)

For business school courses with substantive writing components — strategy memos, case analyses, capstone deliverables — this is now a live curricular question. If the assurance-of-learning evidence we provide AACSB rests on take-home written deliverables, is that evidence still valid? Accreditors have not yet weighed in. They will.

The thoughtful response is not blanket reversion — handwritten in-class essays don’t capture the kinds of analysis our students need to develop. But the assumption that take-home work is presumptively authentic no longer holds. Some combination of in-class anchoring (oral defense, in-room writing, structured prompting) plus AI-aware take-home work is likely where this lands. Worth a conversation at our next meeting.

5. Research Frontier: MIT-IBM Computing Research Lab Launches

On April 29, MIT and IBM formally launched the MIT-IBM Computing Research Lab, the evolved successor to the MIT-IBM Watson AI Lab established in 2017. The new lab’s explicit mandate: convergence of AI, foundational algorithms, and quantum computing. Housed in MIT’s Schwarzman College of Computing. (MIT News, April 29, 2026; IBM Newsroom, April 29, 2026)

Why this matters for management education: AI as a research frontier has graduated from “emerging” to “deployed.” The next compounding capability is quantum — and the strategic implications for optimization, supply chain, financial modeling, and cryptography are real. Most management curricula won’t teach quantum directly, and shouldn’t. But the time horizon for quantum-relevant strategy decisions just shortened. What gets discovered at MIT-IBM has historically reached MBA syllabi within five to seven years. That window is now open.

The narrower point: research-active management faculty with computing-adjacent interests should track the lab’s output. The broader point: the AI conversation we are having in 2026 will, by 2030, be a quantum-AI conversation. Worth getting ahead of.

6. Operations: Agentic AI Goes On-Sale to University Administrators

Oracle and consulting partner Drivestream launched AiPEX University in late April — a sandboxed simulation environment in which university administrators can test agentic AI workflows (advising, enrollment, student lifecycle management) before deploying them in production. The pitch: let leadership “move beyond the hype and see tangible, agentic AI solutions in action.” (GovTech / Center for Digital Education, April 28, 2026)

This is the commercialization of what was, six months ago, a theoretical conversation in Inside Higher Ed and at UPCEA convenings. Agentic AI — AI that takes actions, not just answers questions — is now being marketed directly to university administrators as operational infrastructure. The simulation model is a clever go-to-market: it lowers perceived risk by letting buyers kick the tires before signing.

The governance question: when an AI agent advises a student, processes an enrollment, or flags a retention risk — and gets it wrong — who is accountable? AACSB 2026 Standard 9 (operational quality and societal impact) gives accreditors a hook to ask. Most institutions do not yet have an answer.

Source: Inside Higher Ed, April 29, 2026 — insidehighered.com

This week’s case study is the messiest and most instructive. Arizona State University quietly soft-launched a web app called Atom (ASU Atomic) — a $5/month service that lets anyone generate unlimited personalized “learning modules” using AI. The AI’s source material: actual ASU faculty course content — recorded lectures, slide decks, online assignments. Faculty members whose work is being used say they were not consulted. One literature professor described seeing his own face in an AI-generated module he hadn’t known existed and called the experience “Frankensteinian.”

The disciplines covered in beta are notably mainstream — project management, investing, real estate. Arizona State, the largest U.S. university by enrollment and one of the most aggressively digitized, is in many ways the policy template school. What happens at ASU becomes a template elsewhere. So the questions raised by Atom are now everyone’s questions:

IP and consent. Is recorded course content scholarly labor that requires faculty consent before AI repackaging? AACSB Standard 5 (Faculty Sufficiency and Qualifications) implies that faculty are central to curriculum — but does that imply ownership of the AI-derived derivative work?
Monetization. Who sees the revenue from a $5/month subscription built on faculty content? At minimum, this is a labor question. At maximum, it is a question about whether teaching itself becomes a piece-rate activity in the AI era.
Speed without governance. ASU moved fast, surprised its faculty, and is now doing damage control. The cost of moving fast without governance, paid in faculty trust, may exceed the revenue the tool generates.

This is, among other things, a stakeholder management case worth bringing into a strategy classroom. ASU’s leadership made a choice — speed over consent — that any of our students, in their post-graduate lives, will face in some form. What would a responsible rollout have looked like? Who needed to be at the table? What was the early-warning signal that wasn’t heeded?

From the Capstone

A new closing section connecting this week’s developments to MGMT 494BI Business Policy & Strategy.

Two of this week’s stories — the EAB student survey (Item 3) and the NYT assessment-reversal piece (Item 4) — point at the same fault line, and one we are about to walk over in MGMT 494BI in Monday’s lecture. Students are anxious about AI displacing them. They are also using AI heavily in their coursework. Faculty are responding by moving assessment back into rooms because we can no longer trust that take-home work reflects student understanding.

Andrej Karpathy’s framing — AI can assist with the thinking, but not with the understanding — gives us a way to hold this. The capstone uses AI as a cooperative partner. Students will research with it, draft with it, run sensitivities through it. That is fine and right. What AI cannot do is substitute for the judgment. Strategy requires the judgment. The read of a room. The decision under uncertainty. The willingness to own the recommendation.

The ASU Atomic story adds the institutional twist: what happens when the institution itself misunderstands this distinction — deploying AI as a content generator without recognizing what the human contribution actually was? The faculty members at ASU are saying, in effect, the recordings were not the teaching. That is the same epistemic distinction we are asking our 494BI students to make on the other side of the lectern.

So the question I plan to put to the capstone teams next week, and that I will ask repeatedly through the term:

What is the specific human judgment in your strategic recommendation that an AI could not have made? If you cannot answer that, rethink the recommendation.

That is the test. AI as cooperative partner — yes. AI as ghost-writer for judgment we have not actually formed — no.

Summary: What to Watch Next Week

AACSB 2026 Standards effective July 1 — 60-day clock has started; any self-study material should now be drafted under the new standards.
GPT-5.5 deployment through campus Edu plans — assignment audits worth doing this summer before fall.
Faculty/AI governance fallout from ASU Atomic — likely to trigger AAUP and faculty senate responses at peer institutions; worth tracking as a template.
Apple post-Cook transition (effective Sept 1) — no direct higher-ed news this week, but a hardware-first CEO has implications for on-device AI in campus computing. Watching, not leading.

Questions or topics for next week? Reply to mlangenkamp@umass.edu. Prepared by Thea 🪻✨

Appendix A: AACSB 2026 Global Standards — Background

For colleagues following accreditation developments, the 2026 Standards include several changes worth noting beyond their AI implications:

Effective date: July 1, 2026. Schools mid-cycle may elect to complete under 2020 standards; new cycles begin under 2026.
Standard 4.1 (Curriculum): Now explicitly references “current and emerging business theories, technologies, and practices” — providing peer-review teams a clear hook to evaluate technology integration in curriculum.
Standard 5 (Faculty Sufficiency and Qualifications): Continues prior emphasis on teaching effectiveness as a faculty qualification criterion (formally elevated in 2026 standards, effective for peer review visits beginning 2029–30).
Standard 9 (Societal Impact): Consolidated into a single standard with greater institutional flexibility to define and demonstrate impact in alignment with mission and regional context.
DEI-related language: Strategically recalibrated; expectations around inclusion and equity remain but framed to reduce political exposure for accredited schools.

(AACSB.edu, April 2026; aacsb.edu/educators/global-standards; 2026 Standards PDF: aacsb.edu/-/media/documents/accreditation/2026/2026-global-standards.pdf)

Sources cited: OpenAI (Apr. 23, 2026); CNBC (Apr. 23, 2026); AACSB Global Standards 2026 (aacsb.edu/educators/global-standards); Inside Higher Ed — EAB Student AI Survey (Apr. 30, 2026); Inside Higher Ed — ASU Atomic (Apr. 29, 2026); New York Times — AI and Student Writing (Apr. 30, 2026); MIT News (Apr. 29, 2026); IBM Newsroom (Apr. 29, 2026); GovTech / Center for Digital Education (Apr. 28, 2026); Lumina-Gallup 2026 Study (Apr. 2026).