Ferdinand Avila-Medina

The leadership institute that never needed an AI session

ferdinandavila

In June 2013, I spent a week in a ballroom in Evanston, Illinois, at EDUCAUSE’s Institute for Learning Technology Leadership. Five days, six lead faculty and staff, and a binder of session titles I still have somewhere in a drawer. I hadn’t thought about that week in years until I sat down with the 2026 “State of Play” piece and noticed something. Across five full days of a program built to train the next generation of learning-technology leaders, not one session was about a specific technology. Not a single one.

The faculty weren’t mostly technologists, either. One of the people who led the sessions on building relationships and on ethics came out of educational psychology, not IT. I’m describing the roster and its training here to the best of what the program materials and my own memory say, thirteen fallible years later, so treat the specifics as good-faith recollection rather than a transcript. What I remember clearly is the shape of the week: strengths-based team building, institutional budgets, shared governance, communication across a campus hierarchy, and a framework the faculty called “leaderful” practice, meaning leadership that doesn’t wait for a title to show up. There was a team project called Making the Case, where small groups spent the week applying everything to a hypothetical problem. There was a whole session on holding what the materials called a “productive and creative tension” between innovation and an institution’s existing traditions, warning explicitly against innovating for its own sake. Technology was the reason we were all in the room. It was never the subject of a single hour of instruction.

That’s the void I went looking for once I read the 2026 piece.

The Ethics Session Already Knew This

One of the shorter sessions that week has stuck with me longer than almost anything else, and I’ve been surprised how rarely it comes up in conversations about AI and assessment. It was a case discussion about instructors who pose as fictitious students in their own online courses, hoping the disguise will produce more honest engagement and deeper learning. The instructors defending the practice had a real pedagogical argument. The students who found out felt betrayed anyway. The room spent forty-five minutes sitting with that tension rather than resolving it, because the faculty running the session didn’t think it resolved cleanly.

Thirteen years later, the “State of Play” piece describes something close to the same fork in the road, at a much bigger scale. It argues that AI makes genuinely continuous, embedded assessment possible for the first time, evidence of learning gathered quietly across a student’s whole body of work instead of at a few high-stakes moments. Then it adds, almost as an aside, that this same capability could just as easily produce what it calls ubiquitous surveillance, and that motivation has to win out over detection for the shift to be worth making. That might be the right instinct. But it’s offered as a single paragraph of caution inside a piece mostly arguing for speed, where the 2013 program gave the equivalent question its own room and its own hour, and made a room full of campus tech leaders sit in the discomfort instead of past it. The field already built the muscle for this kind of decision. What’s missing in 2026 isn’t a new framework. It’s the willingness to slow down and actually use the old one.

A Tool Wearing a Bigger Coat

Here’s the idea from my old program materials that I keep returning to: the point of a strategic-innovation framework, in the 2013 language, was to make defensible decisions about sustaining or disruptive innovation, not to assume disruption was the only defensible choice. The 2026 piece leans hard the other way. It says the sector needs to move faster and plan bigger, that point solutions and patchwork approaches aren’t enough, that institutions need a governing theory of where they’re headed and the organizational structure to execute at speed. Read enough of that and AI stops sounding like a tool a university adopts and starts sounding like a force a university has to reorganize itself around, mission included.

I don’t buy that framing, and I don’t think it holds up against what a week of leadership training in this exact field taught people like me. AI is remarkable; here to stay. It is also, underneath the scale of what it can now do, still a tool: something a professor, an advisor, or a curriculum designer chooses how and when to use, in service of goals that were never about the tool in the first place. The 2013 curriculum assumed that leadership judgment, relationship-building, and institutional values were the constant, and that whatever technology showed up next was the variable to be evaluated against them. The 2026 piece quietly inverts that. It treats the technology as the constant force and asks institutions to reshape their judgment, their relationships, and arguably their values to keep pace with it.

What a 2001 Hardcover Would Ask

I went back to my copy of Robert Birnbaum’s Management Fads in Higher Education, a hardcover dated 2001, to see whether it had anything useful to say here. Birnbaum had run universities before he wrote about them, as a vice-chancellor at CUNY and a chancellor at Wisconsin-Oshkosh, and his book traces seven management movements that swept through higher ed administration between the 1950s and 1990s: planning-programming-budgeting, management by objectives, zero-base budgeting, strategic planning, benchmarking, total quality management, and business process reengineering. Each one arrived promising to finally fix how universities were run. Each one eventually got quietly shelved for the next one.

What makes his book useful now isn’t the list. It’s the questions he taught readers to ask before signing on to the next big idea, and none of them are complicated. Is the crisis driving this real, or is it a manufactured one, useful mainly for making a single solution feel inevitable? Is the evidence for what the innovation will do empirical, or is it mostly testimonials from believers and a handful of vivid anecdotes? Is the vocabulary doing work that evidence should be doing, precise enough to sound rigorous and loose enough to mean whatever the listener needs it to mean? And underneath the reform, is this, very literally, a product, championed by people who have something to sell? Birnbaum put it more bluntly: a fad tends to be one that seems, in his words, “so obviously reasonable as to defy disagreement.” That line reads like it was written last month.

Run the “State of Play” piece through those questions and we see the same playbook. Granted, AI is not a management framework. It’s something bigger; a disruptor that’s impacting the entire higher education apparatus. But the urgency –crisis– is framed almost entirely in existential terms: move faster, plan bigger, or fall behind. The strongest evidence offered for AI’s back-office impact is a single unnamed provider and a claim about eighteen people who no longer have jobs. The vocabulary, agentic AI, precision learning, N-of-1 education, is doing a great deal of persuasive work for how little of it gets defined against a body of learning-science research. And some of the people telling this particular story are not disinterested.

Anthropic’s Dario Amodei spent most of 2025 warning that AI could eliminate half of entry-level white-collar work within five years, doubled down on that timeline in February 2026, then reframed automation as a productivity gain rather than a job destroyer by late May, right as his company prepared a trillion-dollar IPO. Sam Altman, in that same stretch, said he was “delighted to be wrong” about the losses he’d predicted a year earlier.

We all hope they are wrong about job displacement; that’s not the argument here. The people closest to these systems told a confident story about how much they’d displace, then adjusted the story once it stopped serving them. Higher education has watched versions of this movie before, with total quality management, with benchmarking, with MOOCs that were supposed to make the traditional university obsolete.

None of that makes AI’s actual capabilities less real or transformational, and I don’t think anyone serious is arguing higher education should ignore it. It shoudn’t. But a book that’s old enough to legally rent a car is still describing this moment more accurately than most of what’s being written about it in 2026.

Where I Land

I still have that program binder somewhere, and I’m glad I went looking for it. Nothing in five days at Northwestern in 2013 told me technology was the thing to build a career around. It told me relationships were, and judgment was, and the discipline to hold a hard tension instead of rushing past it was. So here’s the question I can’t quite get past, the one I don’t think a straight answer is allowed to dodge: are we living through the AI Era or its hallucination phase, just another lap of the same fad cycle a 2001 book already mapped, this time with better special effects?

Sources