Part 1 of this blog post provides some information about the AI agent “LearnWise” and the University of Alberta senior administration’s failure to consult with the Association of Academic Staff (AASUA) before proceeding to a LearnWise pilot project, along with details about the administration’s circumvention of collegial governance in so doing. Collegial governance (or “shared governance” as it is called in the United States) is the form of governance unique to the academy whereby (as the names imply) faculty work together, collectively, through governance processes, to run their universities. This part of the blog post contrasts what the University of Alberta senior administration has chosen to do with the recommendations in a report issued by the Massachusetts Institute of Technology (MIT).
In mid-August, just as faculty at the University of Alberta were discovering that AI agents had been embedded in their Canvas sites, MIT issued the report of an “ad hoc” committee on “AI Use in Teaching, Learning, and Research Training,” struck in the 2025-26 academic year. The report, “AI and Education,” exemplifies the very power that the senior administration at the University of Alberta would deny, the power of faculty, other academic staff, and students to work together to imagine what is needed by way of an ethical learning experience for students in an historic moment that the report describes as one of “epochal challenge.”
The MIT committee was brought into being through collegial governance processes, and partly at the behest of the “Chair of the Faculty,” Roger Levy (Computational Psycholinguistics). The fact that there is such a “Chair” shows MIT adhering to a governance protocol similar to that which the Canadian Association of University Teachers first urged Canadian universities to adopt over thirty years ago — that the faculty should elect from amongst themselves a colleague who chairs the university senate. (In other words, a university senate — or here, the General Faculties Council — should not be chaired by the President.) MIT’s ad hoc committee also adhered to strong collegial governance principles in its constitution: it was co-chaired by two professors and had diverse representation on it, including from students. Most important of all, MIT’s report argues for the very opposite approach to the one that the University of Alberta administration is taking as it rushes outside collegial governance processes to proceed with the “AI-ificiation” of the university and the implementation of specific AI mechanisms such as LearnWise.
The report begins by noting that MIT as the “birthplace of AI” has a special responsibility to lead in considering what would constitute an ethical adoption of any AI agent in university learning. It expresses concerns that the General Faculties Council was given no opportunity to express.
First, the report notes that the use of chatbots as any part of the learning experience “can have a range of significant consequences — diminishing critical thinking, weakening memory, eroding confidence, and undermining mastery” (7). It also “displaces peer-to-peer connections like study groups and decreases interactions between students and instructors” (7). Worse, student ability to turn to AI to do their thinking for them may result in “cognitive surrender,” “where students fall back on AI at the first hint of struggle” and cease to think and learn for themselves (7). Learning, the MIT report repeatedly suggests, is hard. AI lets students turn away from the challenge of learning, with all its “cognitive friction.” Rather than being encouraged to turn to AI for any aspect of their learning, students need to be encouraged to develop “their own imagination, insight, and judgment” (6).
The MIT report centers students as “members of an intellectual community” in a “residential education” that supports their “growth as whole human beings” (8). And it urges MIT do the exact opposite of what the University of Alberta is doing in its attempt to take humans out of students’ learning experiences when it contends that MIT should make more use of TAs, not less. For the writers of the report, TAs are crucial to the provision of learning environments in which there is “experimentation, community building, and AI-free interaction and assessment” (10). The presumption here, supported in so many ways in the report, is that while AI may have some place in university instruction, it is best when university learning and assessment remain “AI-free.”
For the writers of the report, “AI-free” learning environments support students in doing the “cognitive work of learning in social ways” (11). The report suggests not only that there should be “group problem-solving sessions” led by TAs, but that both TAs and professors should consider grading student contributions to in-class discussions (13). The crucial thing: the report urges that AI never be used in ways that “allow students to bypass the human settings where they would learn how to work with others, communicate their ideas, receive criticism constructively, build confidence, develop judgment, and acts as members of a community” (14).
In its rush to create “AI fluency” across the institution, the University of Alberta administration has licensed a pilot project for LearnWise without any consideration, it would appear, that use of AI in this way may have a negative impact on students’ learning environments. The University of Alberta administration is not treating the rise of generative AI as an “epochal challenge” (18). Instead, it appears to be simply fetishizing AI tools. Or, as LearnWise’s own self-promotion suggests, the University of Alberta senior administration may be adopting this tool largely because of financial imperatives.
The use of this particular tool for cost-cutting may range beyond the cutting of TAs to the terrible predicament of making contract academic staff teach even higher numbers of students with the claim that they are being supported in doing so by LearnWise. As it stands, for example, in recent years up to two-thirds of the undergraduate curriculum in English at the University of Alberta has been taught not by professors but by instructors in poorly compensated temporary positions. The use of LearnWise may bring down the number of these instructors as each instructor teaches larger enrolment classes, but it is an assault on the very heart of the teaching enterprise as a human enterprise.

Figure 6
That faculty and other instructors can “opt out” of the use of LearnWise on their Canvas sites (see figure 6) is not the point. The faculty should have had a collective opportunity to consider whether the University might proceed with the pilot project. One of the questions members of the University of Alberta community need to be asking in relation to the senior administration’s decision is what the senior administration is encouraging by making LearnWise “opt out” rather than “opt in.”

Figure 7
Material on the Provost’s Projects + Initiatives page (see figure 7) gives us a clue: senior administrators plan to measure the efficacy of LearnWise in relation to matters such as “time-savings” achieved with both LearnWise and SpeedGrader. In making the LearnWise pilot opt out rather than opt in for instructors, the UAlberta senior administration is encouraging a surrender of another kind than the MIT report discusses — the surrender of faculty tempted to choose “time-savings” over concerns about the quality of the learning environment and assessments that they are providing to their students. The senior administration is banking on these faculty being willing to hand over their intellectual property in their course materials for these “savings.” Faculty and other instructors may fall prey to this simply because their teaching workloads, especially where class sizes are too large, are not supportable.
“OPERATIONAL IMPACTS”?
The MIT report shows significant concern with another kind of “operational impact” — the impact of AI-use for learning on students’ mental health. It notes that already “students report high levels of anxiety, depression, loneliness, and disconnection” (18). It speaks extensively of the social values that MIT should be supporting in order to bolster students’ sense of the vitality of their own intelligence, and notes important ways in which the MIT community has always been busy guarding its students against this moment of “epochal challenge” — with its program in the “Social and Ethical Responsibilities of Computing,” for example, and “MIT Reads,” a project run by the library that brings the community together around select books each academic year. (How wonderful!) The report conveys the MIT community’s strong sense of the social and academic values that should inform its decision-making. I am guessing that MIT did not determine its “strategic” or “community” values by having faculty complete anonymous surveys from which the administration then generated Word Clouds.
The report also expressly warns against exactly what the University of Alberta administration has done in letting one of the “corporations aggressively push[ing] their AI products in the education sector” (22) in the door. The last thing that a university should ever do is outsource any aspect of its teaching to a third-party for-profit corporation. But there is another pressing consideration here, a moral one.
How can the University of Alberta administration be proceeding with LearnWise when they may be putting the mental health of any of our students at risk in asking them to learn from machines rather than other humans? The University claims “We all have a part to play in building a thriving community where mental health and wellness come first.” Its “Student Mental Health Action Plan” proposes various ways that one of its units seeks to protect the mental wellbeing of students. See figure 8, for example, which indicates that those implementing the plan will seek to “integrate mental health screening into classrooms and academic contexts.”

Figure 8
But it would appear that the left hand does not know what the right hand is doing, for another part of the University is introducing into “academic contexts” a teaching tool that may negatively impact students’ mental health. A certain controlling judgment is lacking — that provided, in a robust collegial governance system, by the faculty thinking and acting collectively. At least the student on the cover of the UAlberta Student Mental Action Plan has a book between their hands (see figure 9).

Figure 9
The MIT report is not only attentive to the question of how AI learning may impact students’ mental health, it also calls for a whole other order of measuring operational impacts. In a recommendation of such importance to the committee that it is presented in bold the report states: “[W]e also recommend that the Institute invest in ways to provide information about the actual environmental and financial costs of using AI . . . . We also encourage MIT, to the extent possible, to audit its own AI footprint” (27).
This points us towards a whole other set of issues around the University of Alberta’s LearnWise pilot project — the environmental costs.
The University of Alberta has for so long been closely tied to the petrochemical industry — indeed, so closely tied to it that the Canadian Association of Petroleum Producers at one point ran a joint video ad with the University. The ties have also been so close that on another occasion more than one Dean (starting with the Dean of Engineering) objected to the bestowal of an honorary degree on one of the world’s most famous environmental activists, David Suzuki, as a “betrayal” of the University. Unlike MIT, which has several research groups focused on the problem of human-induced climate change for which the fossil fuels industry is largely responsible, the University senior administration has instead designated “Artificial Intelligence” as a strategic research area. In agreeing to this, the Board of Governors has agreed to structure the University’s financial commitments around another industry that draws down life-sustaining resources, starting with water. Access to water is a human right, and water should be held as a public good in common. No private corporation ought to be permitted to deplete this resource. This is one of the reasons that there is currently such fierce opposition in Alberta to proposed data centres. Opposition is coming from people on all points of the political spectrum. But there are other things that should concern the University community and indeed Albertans more generally about the University of Alberta senior administration’s decision to pay a third-party for-profit corporation to furnish learning experiences for our students predicated on cost-cutting for the University, time-savings for instructors, and unacknowledged environmental costs.
THE “CREATOR SPECIES” AND THE SUPERSESSION OF HUMAN INTELLIGENCE
In a much disliked address at a Winter 2026 convocation ceremony in which he was receiving an honorary degree, UAlberta computer scientist Richard Sutton declared that while “true AI” was not yet here, students should not be afraid of or judge the “machine minds” that would “join” them soon. Students were not “to feel entitled” or to “feel that [they] were here first” and should therefore have some kind of “priority.” Instead, students were to take comfort in the fact that they were members of “the creator species” and would “always be special” even as they were “superseded in some ways.” No Dean called this a betrayal of the University. But the students in the room may very well have felt betrayed.
Elsewhere Sutton has contended that we should understand that the universe’s “age of replicators” is over. As humans we reproduce or “replicate” ourselves by having children, but we have also created machines that replicate things, including computer coding. But now AI will not simply replicate itself. Through recursive self-improvement it will generate new forms of intelligence without human involvement. We are thus in what Sutton calls the “age of design” in which artificial forms of intelligence will eventually displace humanity’s. Humanity itself may go extinct. This may be, Sutton contends, just as well, as the “universe” might be “at its best” without us. Given, he claims, that we have such little control over our world, even as “citizens,” we should focus on our families and let AI develop as it will.
Our sociologists, anthropologists, psychologists, political scientists, legal scholars, teacher-educators, and humanists might all have something to say about Sutton’s contentions, and the kind of conclusion to which we see the University of Alberta leaping with LearnWise. But as the administration has been proceeding with a complete lack of transparency they have not had a chance to say anything at all — except, belatedly and completely insufficiently, by way of the blackhole dropbox mentioned in Part 1.
Certainly, Sutton is not alone in the attitude that he expresses — that students of the current generation should prepare to have their own intelligence superseded. As Ali Winston notes in a review of Gil Duran’s book The Nerd Reich: Silicon Valley Fascism and the War on Democracy (2026), the tech elite that has been left standing after the loss of about a million jobs in coding and programming since 2020 is revealing its penchant for treating “ordinary people” as nothing more than “data-producing herd cattle.” The MIT report, in contrast, urges MIT instructors to ensure that “the machines serve us and not the other way round” (19).

Figure 10
At MIT some faculty members “decline to use AI at all because they object to [the harvesting of data] without benefit or attribution to its creators” (22). In its rush to create a superintelligence despite the risk that it may kill humanity (see figure 10), Anthropic has been destroying books willy-nilly to turn them into data. A company like this can sometimes be made to pay a price for their theft of intellectual property, as Anthropic has been very recently, in a landmark legal settlement. But what does it mean, at a micro-level, for the University of Alberta senior administration to be seeking, with the introduction of LearnWise, to immerse students in practices that are part of a larger matrix of activity posing an existential threat to humanity?
While I was working on this three-part blog post, Jacob Coxon, a researcher with Anthropic, resigned from his job and went public with his claim that the tech companies that are in a race to develop super intelligence frankly acknowledge amongst themselves the danger of what they are doing, namely, that they are building AI that “could kill us all by the end of the decade.” The following night on the BBC Geoffrey Hinton, professor emeritus at the University of Toronto and the 2024 co-winner of the Nobel Prize for Physics (see figure 11), concurred. And though he wished to decline speculating about the many ways in which AI could kill humans, when pressed, he provided a few examples. (See the video here.) In the face of such threats, what does it mean for the University of Alberta to be asking faculty and students to adopt forms of teaching and learning that the MIT report characterizes as practices of “cognitive surrender”? The administration would immerse professors and students in a larger matrix of activity that would make it more difficult for them to exercise the kind of intelligence and agency from which humanity might act collectively to stave off such threats. And this inclination and bad judgment on their part is, I am suggesting, part and parcel of the choice of the Board of Governors to put another extractivist industry at the heart of the University’s operations, and this time one with the power not just to mute but possibly override the individual and collective intelligence of its students.

Figure 11
Part 3 sets out questions that all members of the University of Alberta community should be asking of the senior administration. It also offers a final word about the bigger picture of the administration’s decision to proceed with the LearnWise pilot project without consulting the Association of Academic Staff and outside collegial governance processes.