What Are Mark Carrigan's Views on AI in Higher Education?

Mark Carrigan treats AI in higher education as a strategic problem rather than a tooling question. Across his blog and his open-access book The AI Crisis in Higher Education: A Strategic Proposal, his consistent argument is that universities are integrating AI faster than they can understand its effects, and that the risks fall hardest on assessment integrity and on students' foundational knowledge. If you want his position in one line: adopt AI deliberately, protect what it can quietly erode, and stop treating prompting as the main thing students need to learn.

The core argument: speed is the risk

Carrigan's recurring framing is that the danger is not AI itself but the pace of adoption. One post title states it directly — "The risks of integrating AI too rapidly into higher education" — and another, "The all-you-can-eat AI buffet has closed," signals his view that the early period of unrestricted, uncritical access is ending. His book title, The AI Crisis in Higher Education: A Strategic Proposal, makes the same move: crisis first, then strategy, not the reverse.

This matters if you are writing policy, designing a course, or deciding how your institution should respond. Carrigan's position implies that "should we use AI?" is the wrong first question. The better one is "at what speed, with what safeguards, and toward what end?"

Assessment integrity and student judgement

Two of his posts point at the assessment problem from different angles:

• "The two phases of the AI and assessment integrity in literature" — suggesting the debate itself has moved through distinct stages rather than settling.
• "What are the judgements which students are making as they use AI?" — shifting attention from detection to the reasoning students apply when they decide whether and how to use a tool.

The practical implication: integrity policy built only on catching misuse misses the more interesting question of how students are actually deciding. If you work in teaching or policy, that reframing is worth taking seriously — it points toward teaching judgement rather than policing outputs.

Low-tech and no-tech pedagogy

Carrigan proposes "low tech, no tech pedagogy to protect foundational knowledge from cognitive outsourcing." The concern is specific: if a model supplies the thinking that a foundation is supposed to build, the foundation never forms. His proposed response is not banning AI everywhere but deliberately preserving some learning as unassisted.

For a course designer, that translates into a concrete question per learning outcome: is this something the student must be able to do without the tool, or something they should be able to do with it? Carrigan's writing pushes toward answering that honestly rather than defaulting to AI everywhere.

Prompting is not the lesson

"Why prompting should not be our focus when teaching students about AI" is one of his clearest pedagogical claims. Teaching prompt technique, in his framing, trains students to operate a tool rather than to understand what it does to their thinking, their writing, and their field. Related posts — "Using LLMs to help you write papers doesn't help you get them published" and "Don't be a meat proxy for an LLM" — extend the same scepticism to scholarly work.

Access, rationing, and institutional power

Carrigan also writes about the institutional layer:

• "How will universities ration internal access to LLMs?" — treating access as a distribution problem, not just a licensing one.
• "Why AI vendors are targeting public institutions" — pointing at the commercial dynamics behind campus adoption.
• "How visible are you in the LLM?" — raising the question of whose work and presence the models surface.

Together these suggest his view that AI adoption in universities is also a question of who gets access, who pays, and who is made visible — not only a pedagogical one.

Adjacent themes worth knowing

His blog ranges beyond AI: scholarly publishing and "enshittification," the future of the university, productivity and writing (including an interview about Omnifocus), and music writing on techno and acid techno. The AI-in-higher-education thread is the dominant one, but it sits inside a broader interest in how academic work and institutions are changing.

How to use this

If you are looking for Carrigan's position to cite or act on:

• For policy: start from his speed-and-strategy framing, then his access-rationing question.
• For teaching: start from low-tech/no-tech pedagogy and the argument against prompting as curriculum.
• For assessment: start from the student-judgement reframing rather than detection.

His book The AI Crisis in Higher Education: A Strategic Proposal is described on the site as open access, which makes it the natural next step if you want the full argument rather than the blog-level version.