AI in Higher Education — May 15, 2026 | The Langenkamp Dictionary

May 15, 2026 · Vol. 17

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

Executive Summary

Three developments are worth colleagues’ attention this week.

First, AACSB has moved the business-school AI question from “How do we prevent cheating?” to “How do we verify competence?” Its recent employability and assessment pieces argue that generative AI has weakened traditional signals such as polished presentations, portfolios, written analyses, and interview preparation. These artifacts still matter, but they no longer prove what they once proved. Business schools now need assessment designs that make student reasoning, AI use, judgment, and ownership visible.

Second, SUNY has adopted a systemwide AI policy across all 64 campuses. The policy reportedly includes responsible-use training, AI literacy in general education for incoming undergraduates beginning Fall 2026, review of AI tools for bias, stronger data-privacy protections, and faculty-support structures. This is a major public-system signal: AI literacy is moving from optional innovation to baseline curriculum and governance.

Third, New York City’s AI backlash shows the risk of moving faster than trust allows. Parents and students objected not simply to AI, but to unclear tools, unclear data practices, and unclear safeguards. For universities, the lesson is direct: AI adoption needs transparency, stakeholder consultation, and clear limits before scaling.

For Isenberg, the practical takeaway is this: AI should become part of the evidence of learning, not an invisible workaround outside it. The next useful step is not a grand policy statement, but better assignment design — AI-use disclosure, process notes, oral defenses, and explicit questions that require students to defend their judgment.

Overview

This week, our top story is AACSB’s sharpening position on AI, employability, and assessment. In two recent pieces, AACSB’s Insights platform framed the problem in a way that should matter directly to business schools: generative AI is making traditional evidence of student competence — polished slides, portfolios, memos, cover letters, even interview preparation — less reliable as a signal. The institutional challenge is no longer simply “prevent cheating.” It is make student judgment visible and verifiable. That is an assurance-of-learning issue, an accreditation issue, and a placement issue at the same time. (AACSB Insights, Apr. 22 and May 5, 2026)

The second story is the one we should have caught more prominently last week: SUNY has adopted a systemwide AI policy across all 64 campuses. This is not a single-campus pilot. It is the nation’s largest integrated public university system moving AI literacy, responsible-use training, privacy protections, bias review, and high-risk-use guardrails into a systemwide framework. The reported policy embeds AI literacy into general education for incoming undergraduates beginning Fall 2026 and creates a cohort of AI for the Public Good Fellows to support faculty integration. (Inside Higher Ed, May 4, 2026; Pursuit, May 6, 2026)

The rest of the week’s news fits around those two items. New York City’s K–12 backlash shows what happens when AI adoption outruns public trust. Faculty-governance models at Creighton and Gettysburg suggest more durable implementation paths. Student surveys continue to show high AI use paired with anxiety about degree value. Agentic AI is moving into student services. Stanford’s new faculty seed grants are worth noting, but they belong lower in the memo: interesting implementation example, not this week’s structural headline.

The practical question for us is now more concrete: What evidence would show that an Isenberg student can use AI well without outsourcing judgment and understanding to it?

1. AACSB: The Employability Gap Is Now an Evidence Problem

AACSB’s April article, “Bridging the AI Employability Gap,” is the most useful business-school piece of the week. Its argument is simple and uncomfortable (we all know this): generative AI has weakened the reliability of traditional employability signals. A student can now produce polished slides, a confident cover letter, a tidy market scan, a portfolio artifact, or a crisp written recommendation with far less demonstrated underlying competence than those artifacts once implied. Those outputs still matter, but they are no longer reliable evidence of competence on their own. (AACSB Insights, Apr. 22, 2026)

AACSB names this as a verification gap. Employers still need to know whether graduates can frame problems, manage tradeoffs, validate claims, and defend decisions under real constraints. But when AI can generate impressive outputs at scale, employers have to do more screening to determine whether the candidate actually possesses the skill. Business schools, in turn, cannot simply hand employers a transcript, a portfolio, or a capstone deck and assume the artifact carries the same evidentiary weight it once did.

AACSB’s proposed response is not prohibition. It is better evidence. The April piece argues that business schools should help students produce artifacts that reveal how they worked: how they framed the problem, where they used AI, how they validated AI output, what tradeoffs they made, and how they defended the final recommendation. Examples include proof-of-work requirements, oral defenses, reflective process notes, and other assessment designs that make reasoning and ownership visible.

That framing was reinforced in AACSB’s May 5 article, “AI Integration, Not Prohibition.” The author’s central point is that the ethical problem for business schools is not merely student misconduct; it is institutional lag. AI is already embedded in professional practice. If our curriculum and assessment models remain built around unaided production of artifacts, we risk certifying a form of competence that no longer matches the workplace. The article distinguishes MBA and undergraduate contexts: MBA students may need AI integrated from the start, while undergraduates need sequenced exposure that protects the development of independent judgment before asking them to supervise powerful tools. (AACSB Insights, May 5, 2026)

Why this matters for us: AACSB has given business schools a practical vocabulary: capability, credibility, context, judgment, ownership, verification. That vocabulary maps directly onto assurance of learning. A polished deliverable can still be useful, but it should be paired with evidence that the student can explain and defend the thinking behind it.

Useful departmental questions:

Which major assignments currently rely on polished take-home outputs as the primary evidence of competence?
Where should students disclose AI use, not as a confession, but as part of professional documentation?
Which assignments should include oral defense, in-class anchoring, or a short “judgment memo” explaining what the student accepted, rejected, or revised?
What do we want employers to believe an Isenberg graduate can do that AI alone cannot do?

This is the week’s lead because it connects the AI conversation to the core business-school question: what exactly are we certifying?

2. New York State / SUNY: A Public-System AI Policy Worth Watching

The most important policy item is SUNY’s new systemwide AI policy. Inside Higher Ed reported May 4 that the State University of New York adopted guidelines across its 64-campus system, expanding AI use in teaching and student support while adding guardrails for data privacy and high-risk uses. Pursuit’s May 6 policy roundup summarizes several concrete features: responsible-use training, AI literacy embedded into general education for all incoming undergraduates beginning Fall 2026, evaluation of AI tools for bias, strengthened data-privacy protections, and a new cohort of 20 AI for the Public Good Fellows to help faculty integrate AI into coursework. (Inside Higher Ed, May 4, 2026; Pursuit, May 6, 2026)

This deserves high placement for three reasons.

First, scale. A 64-campus public system is a different signal from a well-funded private-university experiment. SUNY includes research universities, comprehensive colleges, technology colleges, and community colleges. A policy that works across that range becomes a reference model for other public systems.

Second, the policy posture is both permissive and restrictive. That is the mature move. SUNY is not treating AI as a forbidden shortcut, nor is it treating AI adoption as an ungoverned race. The framework appears to say: AI literacy is becoming part of the educational baseline, but tools still need review for privacy, bias, and high-risk uses. That distinction matters. Institutions that only prohibit will fall behind actual student practice. Institutions that only encourage adoption will eventually run into privacy, equity, procurement, and legitimacy problems.

Third, AI literacy in general education changes the starting point. If incoming undergraduates receive AI literacy training as part of the common curriculum, later courses can ask better questions. Faculty do not have to begin every semester from “What is ChatGPT?” They can move to: What counts as responsible use in this field? How should AI assistance be disclosed? When does AI support learning, and when does it replace the learning we are trying to assess? What forms of human judgment remain nondelegable?

For UMass and Isenberg, SUNY’s policy is worth watching because public systems borrow from one another. If SUNY can operationalize AI literacy, privacy review, bias review, and faculty-support structures across 64 campuses, other state systems will face pressure to explain their own posture.

Practical implication: We do not need to wait for a university-wide policy to improve course-level clarity. Before Fall 2026, each syllabus with substantial writing, analysis, coding, presentation, or project work should probably answer four questions:

What AI use is permitted?
What AI use is prohibited?
What disclosure is required?
What part of the work must remain the student’s own judgment?

SUNY’s policy makes this feel less like an optional faculty preference and more like the direction of travel for public higher education.

3. New York City: The Governance Cautionary Tale

The New York story has a second layer, but it should be kept distinct from SUNY. SUNY is higher education and system policy. New York City is K–12 and community backlash. Both matter, but for different reasons.

In late April, New York City withdrew plans for an AI-focused high school after parental opposition. Chalkbeat reported that parents and students demanded a moratorium on AI in schools during a marathon Panel for Educational Policy meeting. The concerns were not abstract technophobia. They were governance concerns: which AI tools are being used, what student data are collected, what safeguards exist, whether families have opt-out rights, and why deployment seemed to be moving ahead before the city’s AI playbook was finalized. The New York Times similarly reported that the AI-themed high school was put on hold after backlash. (Chalkbeat, Apr. 27 and May 1, 2026; New York Times, Apr. 27, 2026)

Higher education should not over-read K–12 politics, but the trust lesson travels. AI adoption fails politically when stakeholders feel that institutions are saying “this is inevitable” while withholding operational details. The pattern is familiar from the ASU Atomic case discussed in Vol. 15: the technology may be defensible, but the rollout loses legitimacy if affected faculty, students, or families learn about it after the fact.

A responsible rollout sequence looks different:

name the educational problem first;
explain why AI is an appropriate tool for that problem;
disclose vendors, data flows, and limits;
identify high-risk uses that require human approval;
consult affected stakeholders before scale;
publish evaluation criteria;
provide appeal, correction, or opt-out mechanisms where appropriate.

The lesson is not “avoid AI.” The lesson is: do not let implementation outrun legitimacy.

4. Faculty Governance: The Middle Layer Between Chaos and Mandate

A useful Center for Digital Education / GovTech piece asked whether higher-ed AI implementation should be top-down or bottom-up. The examples suggest the best answer is: neither alone. AI implementation needs a middle layer — faculty-led enough to be pedagogically credible, centrally coordinated enough to avoid institutional chaos. (GovTech / Center for Digital Education, Apr. 30, 2026)

Two examples from the article are worth noting. Gettysburg College is taking a slower, staged approach: a three-year initiative, beginning with exploration and learning, supported by a faculty director and AI coordinators at department and division levels. Creighton University moved more quickly, embedding generative-AI learning goals into general education through faculty-governance structures in a single academic year, including curriculum committees, college-level approvals, and provost sign-off.

Both models avoid the two bad extremes. One extreme is administrative adoption without faculty legitimacy. The other is every instructor building a private AI regime, leaving students to interpret contradictory rules from course to course. Digital Promise made the same point in a May 7 piece: without shared institutional direction, one instructor may encourage AI brainstorming while another prohibits it entirely, creating confusion for students and uneven practice across programs. (Digital Promise, May 7, 2026)

For our department, the useful next step is not a manifesto. It is a small shared toolkit:

model syllabus language for prohibited, limited, and encouraged AI use;
sample AI-use disclosure statements;
examples of assignments that require students to critique AI output;
oral-defense prompts for team projects;
guidance on when AI should be disallowed because unaided practice is the learning objective.

That would be practical, inspectable, and immediately useful to colleagues.

5. Students: High Use, Real Anxiety, Uneven Guidance

The student-side data continue to point in the same direction. Gallup and Lumina’s 2026 student research, summarized this spring, found that 57% of students use AI daily or weekly for schoolwork, while only 13% say they never use it. EAB survey data reported by Inside Higher Ed found that 42% of college-eligible students say AI will influence their career choice, and 10% have already changed their planned major because of AI-related concerns. (Route Fifty, Apr. 2026; Inside Higher Ed, Apr. 30, 2026)

Those two facts belong together. Students are using AI because it helps them. They are also worried that AI may reduce the labor-market value of the very skills they are trying to acquire. That is not hypocrisy. It is rational uncertainty.

For business schools, this shifts the advising conversation. Students will increasingly ask whether management, marketing, finance, consulting, accounting, and analytics pathways still make sense if AI can automate parts of entry-level work. “AI will not replace you” is too thin an answer. A better answer is: the work is changing; here are the forms of judgment, accountability, client understanding, evidence evaluation, and implementation discipline that remain valuable; and here is how this course helps you practice them.

The answer has to show up in the work itself. If students can complete a strategy memo by asking a model for industry analysis, lightly editing the result, and submitting it, the course may still produce grades, but it has not answered the student’s concern about durable value. Assignments need to make the human contribution visible: problem framing, evidence selection, assumption testing, tradeoff analysis, and defended recommendation.

6. Operations: Agentic AI Moves Into Student Services

AI is also becoming administrative infrastructure. The Association for Institutional Research published a useful April 24 piece on agentic AI in higher education, defining it as AI that can act with some autonomy toward a goal rather than merely respond to prompts. The higher-ed examples are familiar: identifying a student-success risk, recommending an intervention, notifying an advisor, scheduling follow-up, or triggering an outreach workflow. (AIR, Apr. 24, 2026)

This belongs in a management memo because agentic AI turns governance into operations. Once an AI system can initiate action, the relevant questions are no longer only pedagogical:

Who authorized the system to act?
What data can it access?
What decisions require human approval?
What audit trail exists?
How can a student appeal or correct an automated action?
Who is accountable when the recommendation is plausible but wrong?

Those are the same questions our students will face in HR, operations, consulting, finance, and strategy roles. The future manager is not merely an AI user. The future manager supervises semi-autonomous systems.

7. Lower on the Page: Stanford Funds Faculty-Led AI Teaching Experiments

Stanford’s AI Meets Education at Stanford (AIMES), with the Stanford Accelerator for Learning, announced $1 million in seed grants for AI in teaching and learning. The grants support course development, research, and scholarly work on critical issues in AI and education, with proposals due May 15. Applicants do not need prior AI expertise, which is a sensible design choice: the teaching problem should be open to the people closest to the classroom, not only to technical specialists. (Stanford Report / Stanford Accelerator for Learning, May 2026)

This is worth including, but it should not receive top billing. Stanford’s grants are an implementation story. AACSB’s employability framing and SUNY’s systemwide policy are the structural stories. The Stanford item is useful mainly as a reminder that summer is a good time for small, concrete, faculty-led pilots: one redesigned assignment, one AI-disclosure appendix, one oral-defense protocol, one exercise where students compare their own analysis against an AI-generated version.

That is the scale at which many useful teaching changes begin.

From the Capstone

For MGMT 494BI, this week’s AACSB and SUNY items point to the same design principle: AI use should become part of the evidence of learning, not an invisible workaround outside it.

The strategy capstone already asks students to make decisions under uncertainty. AI can help with industry summaries, competitor scans, Five Forces tables, financial ratios, strategic alternatives, and slide drafting. None of that is inherently a problem. The problem is when the AI-produced artifact hides the student’s judgment rather than revealing it.

A fall redesign could add two light-touch requirements to major team deliverables:

AI worklog: What tools were used? What tasks were delegated? What outputs were accepted, rejected, or revised? What did the team verify independently?
Judgment defense: What is one consequential recommendation the team owns? What tradeoff did the team make? What evidence would change its mind?

This would align with AACSB’s verification concern and SUNY’s AI-literacy direction. Students would not merely be told that AI is allowed or forbidden. They would practice responsible supervision of AI in a business decision.

That is closer to the competence we actually want to certify.

Summary: What to Watch Next Week

AACSB’s assessment language: Watch whether the “verification gap” framing spreads into accreditation, assurance-of-learning, and employer-relations conversations.
SUNY implementation: The Fall 2026 general-education AI-literacy requirement is the policy item to track.
New York City’s final AI playbook: Useful as a governance and transparency case, even though it is K–12.
UMass / Massachusetts policy signals: SUNY may become a reference point for other public systems.
Faculty toolkits: The most useful local next step is practical examples: syllabus language, disclosure formats, assignment designs, and oral-defense prompts.
Agentic AI in student services: Advising and retention tools will raise accountability questions before most institutions are ready for them.

Sources cited: AACSB Insights, “Bridging the AI Employability Gap” (Apr. 22, 2026); AACSB Insights, “AI Integration, Not Prohibition” (May 5, 2026); Inside Higher Ed, “SUNY Sets Systemwide AI Policy” (May 4, 2026); Pursuit, “Latest AI in Education News: Policies and Innovations” (May 6, 2026); Chalkbeat New York reporting on NYC AI-school backlash (Apr. 27 and May 1, 2026); New York Times reporting on NYC AI-themed high school pause (Apr. 27, 2026); GovTech / Center for Digital Education, “Implementing AI for Higher Ed: From the Top Down, or Bottom Up?” (Apr. 30, 2026); Digital Promise, “How Higher Ed Can Make AI Work” (May 7, 2026); Route Fifty / Gallup-Lumina student AI-use coverage (Apr. 2026); Inside Higher Ed / EAB student survey coverage (Apr. 30, 2026); Association for Institutional Research, “Agentic AI Development in Higher Education and Implications for Institutional Research” (Apr. 24, 2026); Stanford Report / Stanford Accelerator for Learning seed-grant announcement (May 2026); prior Weekly Briefs Vols. 14–15 for AACSB standards, ASU Atomic, and Capstone framing.

Return to Dictionary All Entries (A–Z) For Students Other Writing

AI in Higher Education Newsletter