My previous Blog showed how AI course companions enhance instruction by bringing affordable 24/7 customized engagement to all students, not just the “visible few” at the top and bottom of the class. They also provide valuable early diagnoses of student learning problems, so professors can intervene in a customized, timely, and effective way. Faculty and students who have climbed the (rather gentle) learning curve to use these Companions first-hand tend to be enthusiastic. Why, then, do many institutions and faculty display viscerally negative reactions to their adoption?
The answer is that the course companions are paradigm-shifting, and thus disruptive, innovations. They upend preexisting workflows and thus threaten those who perform them. Resistance based on such threats is not new. For example, the reaction of major universities was decidedly mixed when Johannes Gutenberg’s movable type made its way across Europe in the mid-to-late 15th century. (Printing can be regarded as a major application of new technology to university teaching.) Some intellectuals rejected the innovation on philosophical, pedagogical, and economic grounds. For example, the Sorbonne’s Theology Department, after having initially embraced the innovation, joined other academic centers in fearing that the wide availability of printed work would destroy academic gatekeeping and textual integrity. The Benedictine abbot of Sponheim (a Monastery) defended his scribes’ livelihoods by arguing that parchment lasted centuries, whereas cheap printed paper was physically fragile and intellectually disposable, creating a dangerous, false sense that knowledge was safely preserved when it was actually fleeting. (He printed his manuscript on paper, however.) Such concerns may sound laughable today, but they were deadly serious at the time.
Creative Disruption
The first element of disruption is change itself: overcoming inertia. Then come threats to perceived self-interest and key values. Such challenges are inevitable. Joseph Schumpeter built his innovation theories around “creative destruction” (a lethal cousin of creative disruption) in his 1942 book, Capitalism, Socialism, and Democracy. The argument’s core was that innovation destroys established interests through the various actions that create new value. Creation of such value is essential. Competition and long-run economic growth will stagnate in the absence of innovation. The established interests will push back; that’s beneficial so long as it generates new creativity and efficiency rather than suppressing the issues. Government’s role is not to insulate people and interests from the adverse effects of innovation, but rather to help mitigate those effects where the free market fails to do so.
What self-interests of academics drive university teaching? Departmental budgets certainly come to mind. Is the innovation intended, fundamentally, to justify budget reductions? Faculty positions and workloads are closely related to budgets. Will the innovation increase workloads? Will the number of faculty positions decline, either directly from the innovation or the increased workload associated with it? Teaching and learning quality is on the table, too. That’s a familiar subject in these blogs. For now, I’ll say only that any innovation that threatens quality is highly suspect in my opinion.
Here is an example of a disruptive innovation in an academic department: one that doesn’t directly impact teaching quality, but rather through departmental budgets and faculty time. It is the introduction of institution-provided personal computers at my university a few decades ago. We take such usage for granted now, but, like the printing press, it was a controversial innovation at the time.
Some faculty welcomed the new PC on their desk. Some thought the idea abhorrent: inhuman, demeaning (“I don’t type”). Some were concerned about their secretaries: will they lose their jobs? Who will keep my calendar and take my phone messages; how will I get my copying done? The university assured faculty that PC usage was the way of the future, and not an excuse to reduce secretarial support. This did not assuage all concerns, but the PCs (and associated training) appeared anyway.
Most readers of this blog won’t be surprised by the outcome. Professors learned to use their computers, and they liked them. Composing academic papers with a word processor was easier than doing so longhand and waiting for a secretary to do the typing. The calendar and messaging apps were quite convenient. The department’s secretarial support remained adequate for copying and other support tasks. There were no layoffs.
Was this another example of layering cost in the name of modernization? It didn’t turn out that way. The annual budget process always had been competitive: not everything the faculty wanted could be funded. It wasn’t long before departments, themselves, decided not to replace some departing secretaries and use the savings for more urgent needs. I’ve seen this happen over and over in various contexts. Everybody wins.
Dynamic Destabilization
Taking a significant innovation seriously destabilizes both individuals and organizations. Established mental models, interpersonal relationships, power structures, and even ways of relating to the world and work are upended by dynamic processes that persist for a considerable period of time.
Clay Christensen’s The Innovator’s Dilemma (1997) elaborates. Disruptive innovations often are cheaper, simpler, or more convenient for some purposes, but worse at the job the current technology was optimized to do. They “serve the underserved” in terms of functionality, price, or both, and their future trajectory is genuinely hard to predict. Disruptive innovations also tend to change rapidly as firms fix problems and search for the right market and functionality. They are a moving target. The constant flux makes it hard for adopters to understand the technology—which undermines the confidence of possible adopters. This dynamic helps explain why incumbent firms, despite having the resources and technical capability, often adopt disruptive innovations too late or not at all.
It’s easy to find fault with an innovation, especially when applied in settings with well-established legacy behavior. The costs, shortfalls, and threats are clear, but future benefits may be hard to demonstrate. No wonder, then, that significantly disruptive innovations may be resisted strongly, on a variety of grounds, in a variety of ways. What couldn’t be known in the mid-2000s, of course, was that within two decades the limited and inflexible learning algorithms embedded in that era’s course objects would be replaced by the power and flexibility of generative AI.
Disruptive Innovations in University Teaching
Gutenberg’s invention of movable type was a quintessential disruptive innovation for university teaching and learning. I believe course companions represent the leading edge of a similar revolution. An earlier blog (Massy, “Technology’s Misunderstood Potential,” GrayDI: July 30, 2026) described the three elements of an algorithm-based Course Object (a precursor of AI-based course companions): (1) a knowledge base; (2) low-friction student access methods; and (3) an event generator for engaging student interest and enhancing learning.
The following maps these ideas to course companions.
- Knowledge Base: instructor-provided, based on the course syllabus and supporting materials. Includes images and symbolic representations. Can use the broader web only with permission.
- Low-friction student access Methods: ability to access the knowledge base using natural language (including a foreign language), in terms of concepts rather than literal lookups. Appropriate guardrails are provided.
- An Event Generator for Engaging Student Interest and Enhancing Learning: mostly the Socratic method, where queries are answered with more queries based on the student’s learning record, although other modes are possible. Moreover, this event generation can alert the professor in real time when a student appears to be performing poorly.
My previous blog, “Customized Learning Engagement for all your Students: at Scale with Low Cost” (GrayDI: August 18, 2026), discussed Socratic dialog’s role in high-quality learning, why it’s so hard to achieve at scale, and how course companions help do this by leveraging faculty effort and engaging students.
It’s instructive to compare course companions to the teaching innovations Clay Christensen studied in his 2011 book, The Innovative University. Online learning provided a clear application of Christensen’s theory. It started by serving the underserved: e.g., by improving on correspondence courses. Faculty and administrators had little basis for predicting how it would affect learning outcomes, completion rates, let alone their workloads or institutional financial health. Delivering high-quality learning economically, at scale, required countless further innovations in content packaging, course structure, and student-faculty interaction. Some of these were made by software suppliers and university staff, but many came from faculty themselves as they applied the new tools to better serve their students. Eventually, online learning was used in on-campus as well as distance courses, and a new genre, the hybrid course, proved effective for students who could visit campus only on a periodic basis.
Nowadays, few doubt that online learning has been worth the investments of time and money that went into its development. Some institutional efforts have failed but, on the whole, the genre has been a great success. Such an outcome is not assured, of course.
Christensen also describes the introduction of MOOCs: “massive open online courses” that allowed unlimited participation on the web. The outcomes so far have been equivocal at best. The early optimism associated with MOOCs dissipated with experience, especially the observed low course completion rates. One reason may have been that generative AI was not available when MOOCs began in the early 2010s: engagement dropped off when students advanced beyond their depth with no one to “talk to.” That’s changing now, however, so the last word on MOOCs has yet to be written.
The generative AI embedded in course companions transforms some of the faculty’s cognitive and behavioral connections with students and technology. This paradigm shift offers immense benefit, but it can also evoke visceral resistance. I’ll explore this resistance and what might be done about it in my next blog.
