AI in Higher Education — April 10, 2026 | The Langenkamp Dictionary

April 10, 2026 · Vol. 12

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

Another dense week in the AI-and-higher-education space. The big story is institutional pushback against large-scale vendor deals — a useful mirror for any department thinking carefully about how AI enters the classroom. Meanwhile, the data on student usage continues to outrun most institutional policy. Five areas covered below.

I. AI Tools for Teaching

Instructure Launches IgniteAI Agent for Canvas (March 23, Inside Higher Ed)

The most consequential LMS development this cycle: Instructure has launched IgniteAI Agent, an agentic AI layer built directly into the Canvas LMS. According to Zach Pendleton, Instructure’s chief architect, the system is designed to handle multi-step instructional tasks autonomously — think automated quiz scaffolding, feedback drafts, and course content suggestions — rather than simply responding to one-off prompts.

Key logistics for Canvas users:

Free access for all U.S. Canvas customers through June 30, 2026 (global access extended through September 30)
The agent can be scoped and constrained by instructors, addressing earlier concerns about AI operating outside faculty control
Instructure describes the architecture as “agentic technology built for education, responsibly, transparently and in partnership with our customers”

Practical implication for Isenberg: If UMass runs Canvas (which it does), this is worth a pilot. The free-access window makes the cost barrier zero. The risk-management question is whether your syllabus and course design are ready for students interacting with an AI layer inside the LMS itself.

Other tools gaining traction in spring 2026:

ChatGPT Edu and Microsoft Copilot remain the dominant instructor-facing tools for lesson planning, rubric generation, and feedback drafting
Khanmigo (Khan Academy’s AI tutor) is seeing expanded university adoption for flipped-classroom models
Perplexity Computer (launched February 2026) orchestrates 19 AI models in parallel and is appearing in graduate-level research workflows

II. Policy & Governance

The CU Boulder Flashpoint: When Faculty Push Back on Vendor Deals

The most instructive governance story of the past two weeks: the University of Colorado Boulder delayed its student rollout of ChatGPT Edu following organized faculty resistance to its $2 million OpenAI contract. (Boulder Reporting Lab, March 29; Inside Higher Ed, March 27)

A faculty open letter, signed by researchers and instructors across departments, cited four core concerns:

Data privacy — Colorado’s public records laws could potentially expose student chat logs to third-party legal requests
Academic integrity — whether institutionalizing AI access implicitly normalizes AI-generated work
Critical thinking degradation — substantive concern that LLM use may undermine the development of analytical reasoning in students
Lack of faculty consultation — the contract was signed before faculty governance bodies were meaningfully engaged

The CU provost issued a statement defending the deal’s data protections, and the rollout was partially delayed rather than cancelled. Faculty achieved a partial victory.

Meanwhile, the California State University System is navigating a parallel dispute around its $17 million OpenAI contract — the largest such deal in public higher education — covering 460,000 students and 63,000 faculty. The contract expires in July 2026, and CSU is deciding whether to renew. A comprehensive system-wide survey (the largest AI study ever conducted in U.S. higher education) found 53% of CSU users engage with AI regularly — but faculty concerns about data handling and pedagogy remain unresolved. (GovTech, Route Fifty, April 2026)

Takeaway for department leaders: The CU and CSU cases are a governance template. If your institution is negotiating or renewing a campus-wide AI vendor contract, faculty need a seat at the table before the ink dries — not afterward. The core questions are not technical; they are about academic values, data sovereignty, and pedagogical philosophy.

III. Pedagogy & Research

What the Data Says About Students and AI in 2026

Two major studies landed this week with complementary findings:

Lumina Foundation-Gallup 2026 State of Higher Education Study (Gallup, April 9)

57% of U.S. college students use AI in their coursework at least weekly
Students who use AI regularly are more likely to report feeling academically prepared — but also more likely to report uncertainty about whether their programs connect to workforce needs
Key policy implication: students want AI skills explicitly mapped to career outcomes, not just embedded invisibly in coursework

HEPI Student Generative AI Survey 2026 (Higher Education Policy Institute, UK)

AI use is now “near universal” among undergraduates in the UK — the question has shifted from whether students use AI to how well they use it
Institutions are behind: most students report receiving little formal guidance on responsible or effective AI use

On the integrity question — some nuance: A Packback survey of nearly 700 U.S. college students found that only ~5% report using AI to generate full assignments — roughly comparable to pre-AI academic dishonesty rates. The majority use AI for brainstorming, outlining, editing, and research synthesis. This does not mean integrity concerns are overblown, but it does suggest the “everyone is cheating” narrative deserves scrutiny.

Springer/Education & Information Technologies (2026): A new peer-reviewed study found that pedagogically guided AI integration — where instructors explicitly frame how and why to use AI tools — produces significantly better active learning outcomes than either AI prohibition or unguided AI access. The implication: the instructor’s role in framing AI use matters enormously.

IV. Workforce Implications

The MBA Curriculum Problem

Two pieces published this month put direct pressure on business schools:

Forbes, March 30: A new survey finds 60% of MBA students believe their programs are outdated for the AI-era workforce. The specific gaps cited: AI-augmented decision-making, prompt engineering for business contexts, AI ethics and governance, and the ability to manage AI-assisted teams. Traditional strengths — stakeholder alignment, cross-functional collaboration, financial modeling — remain valued but are no longer sufficient differentiators.

Poets & Quants, March 19: A student essay argues that business schools don’t need more AI courses — they need an AI-native curriculum. The distinction matters: bolting an AI elective onto a traditional MBA structure produces uneven AI fluency within a cohort, which the author argues is worse than no AI instruction at all. The prescription is integration across every course, not concentration in a few.

DataCamp Workforce Report (2026): 59% of enterprise leaders now report an AI skills gap in their organizations, even among employees who have completed AI training. The finding: training focused on awareness of AI tools does not translate to applied capability in workflows. The implication for business schools is pointed — we need to teach students to work with AI in real business contexts, not just to understand it conceptually.

BCG Research (April 2026): Organizations that treat AI literacy as “core infrastructure” — embedded in workflows, reinforced iteratively, measured against outcomes — see 14–19 percentage point higher AI adoption rates than those offering standalone training. This is essentially an argument for the AI-native curriculum approach.

For Isenberg faculty: The pressure is sharpest in strategy, consulting, and management courses — exactly our wheelhouse. If students are using AI for case analysis, competitive intelligence, and stakeholder presentations, the question is whether we are teaching them to do that well.

V. Case Study: University of Chicago’s Mansueto Faculty Initiative

The Positive Model — Going Wide, Not Just Deep

While much of the news cycle focuses on AI tools for students, the University of Chicago is making a significant bet on the faculty side of the equation. (Forbes, April 3; Inside Higher Ed, April 6; UChicago News)

A $50 million gift from alumni Rika and Joe Mansueto has launched the Mansueto Faculty of Mind and Machine Challenge, a matching initiative targeting $200 million total to recruit and retain 20 leading scholars who integrate computational thinking and AI across a wide range of disciplines — including the arts, social sciences, medicine, business, and law.

This is not an AI department hire. It is a deliberate, institution-wide strategy to embed AI-fluent scholarship across the full intellectual footprint of the university. The explicit goal is to prevent AI expertise from siloing in computer science while the rest of the university watches.

Why this is a model worth studying:

It treats AI as a disciplinary lens, not a stand-alone field
It uses endowed chairs to create permanence rather than relying on short-term grants or vendor relationships
It spans humanities and professional schools — the hardest-to-reach constituencies

The contrast with the cautionary cases: Where CU Boulder and CSU found themselves managing faculty resistance after a vendor deal was signed, UChicago is building buy-in by investing in faculty as the agents of AI integration. The governance model is entirely different — and likely more durable.

For management departments: The lesson is structural. AI adoption at the faculty level requires investment in faculty, not just subscriptions for students. The question for Isenberg leadership is whether our professional development resources reflect this.

Summary & Recommendations

Area	Key Development	Action Item
Tools	Canvas IgniteAI Agent free through June 30	Pilot in one course this spring
Governance	CU/CSU faculty pushback on OpenAI deals	Review UMass’s vendor agreement terms
Pedagogy	Instructor framing drives AI learning outcomes	Develop shared guidance for faculty on framing AI use in syllabi
Workforce	60% of MBA students see curriculum gap	Audit AI integration across core management courses
Case Study	UChicago’s $200M faculty-centered model	Advocate for faculty development investment, not just student access

Questions or topics for next week? Reply to mlangenkamp@umass.edu

Prepared with the assistance of Thea, AI research assistant. Sources: Inside Higher Ed, Gallup/Lumina Foundation, HEPI, Forbes, Poets & Quants, BCG, DataCamp, Boulder Reporting Lab, GovTech, Springer Education & Information Technologies, UChicago News. April 2026.