I believe that one of the underappreciated aspects of the imposition of generative AI technology on educational spaces is the emotional impact on front-line instructors and instructional staff.
For this reason, I’ve been starting my public talks and faculty workshops by simply asking the audience how they’re feeling about this technology. Not how they’re using it, not what effect they think it’s having, but how they’re feeling.
Because of the audiences I speak to—primarily people teaching high school and college or working in direct support of those missions—most are not feeling good. For sure, part of this is the sense that their work has been violently and irrevocably disrupted by generative AI. While the COVID pandemic was surely more disruptive, I think most had the sense that the disruption would ultimately be weathered, and at least for my money, the period of major disruption was much shorter than I would have figured at the onset of that crisis.
With generative AI, there is no such belief. If anything, most believe the degree and manner of disruption will only increase.
There is also a sense that they are largely being thrown to the wolves, sometimes by being at cross-purposes with what they perceive as their administration’s goals, and sometimes because they feel as though they have been left alone, the fact of having a friendly and entertaining sort such as myself to come to talk with them for a day or two aside.
Anger, sadness, even despair are all mentioned. There are also some smatterings of excitement, but again, given the sorts of spaces I’m invited to, this is less prevalent. These emotions are one of the reasons that, from the outset of these talks and workshops, I try to emphasize something I first said less than two weeks after ChatGPT came into the world: We can treat this as an opportunity, not a threat.
The opportunity is not necessarily to leap into a new future where AI is integrated into every fabric of our lives and work. This inevitability is the pitch of the major AI companies, and even if I believed it to be true—which I don’t—it would not benefit me to embrace it because that would be a choice to disempower myself in a world where being maximally agentic is the new advantage.
No, the opportunity is that the capacities of the AI technology allow us to reflect upon and reanimate our human selves in educational spaces that—to me at least—seem to have been moving away from practicing our humanity in the service of fulfilling a rather narrow transaction of credentialing for quite some time, maybe even the entirety of my time as an instructor.
Two recent testimonies from college faculty that I caught on the Bluesky platform drove home how difficult the emotions associated with AI in education and scholarship are, and how it is more necessary than ever to lean into human agency.
One thread came via Crystal Fleming, a professor of Africana studies at Smith College, who related an experience at a digital humanities conference where, as she was taking in a keynote address, she had a nagging sense that the address had been generated using an LLM. During the Q&A she asked the speaker, almost offhandedly, if this was the case, and he admitted that he now worked closely with Claude and was also “writing” a book where the text would be entirely AI-generated.
Professor Fleming reported that the speaker went on to say that using AI led him to feel “useless as an intellectual.”
Personally, I find the fact that this speaker was not planning on disclosing his use of AI without being prompted ethically appalling, but once I got past that reflex, I felt sort of sad for this man. What has happened when a living, breathing, thinking, feeling human being accomplished enough to be invited to deliver a keynote at a conference believes that they can be replaced by a computational process?
Here is someone who has allowed the threat to overwhelm any capacity for seeing opportunity.
In a later part of the thread, Professor Fleming captures the opportunity: “Interestingly, the intuition I felt and spoke from—that embodied knowing—is precisely the kind of thing AI, for all of its performative anthropomorphic gesturing at the simulacrum of personhood, can’t do. Tech billionaires want us to believe that our embodied/human ways of knowing are useless!”
She had experienced proof positive that we are distinct from the computational process, and that the embodied expertise had a capacity that cannot be achieved by a large language model. Clearly, these are the capacities we should be seeking to build in our educational institutions.
The other thread was from Jeff Sharlet, a writer and creative writing professor at Dartmouth who anonymously surveyed his students—to get candid responses—about AI at the end of the semester.
Sharlet’s thread is interesting because it shows a full range of student emotions, most of them some mix of the negative: “Mood ranges from resignation to despair, capitulation from embittered erosion of standards to total, feelings of betrayal from deep to furious.”
Some of the specifics are disturbing—students feeling like they’ve grown dependent on AI, getting worried about such dependence, scaling back, and then seeing that dependence creep back in. Students resented their institution which seems all in on AI (Dartmouth is in partnership with Anthropic) and professors who seem to be using AI to evaluate student work.
At its worst, the testimonies from Professor Sharlet’s students portray a system on the verge of collapse, a mix of cynicism and raw calculation to play the game as they perceive it, rather than centering their own values and growth. In many ways, this is nothing new. The transactional nature of the degree, particularly in elite schools like Dartmouth, has always been not far from top of mind.
But here too is the presentation of opportunity. Professor Sharlet mentions one that I think is obvious, a re-examination of how we assess students and their work by considering alternative grading approaches. In my talks and workshops, I emphasize that this is the most important question to explore. If learning is meant to be the center of the institution, what evidence can we gather to show students are learning?
Our first principle should be to work on making student learning visible to students themselves. This orientation can go a long way toward mitigating those negative emotions around AI use as students see themselves as agentic actors, even inside a system that may seem hostile to their ends.
In my presentations, I try not to give false hope when it comes to navigating the challenges of AI in education, but there is hope to be found. As both testimonies illustrate, we have a strong idea of the work and experiences that will remain human, and we have lots of prior knowledge of how to structure the work and experiences so they remain meaningful.
That keynote speaker once knew how to develop his ideas in the absence of AI bigfooting his own thoughts. Perhaps his confidence has been shaken, but getting back to the work of human scholarship may also serve to reacquaint him with the pleasures of building and sharing knowledge.
Students are pretty clearly expressing what they want and need: experiences that lead to learning. We know how to do this, but we have to make it the focal point of the teaching part of the institution. The institutions that pursue this path will find themselves thriving over time, resilient against whatever further disruption may come.
Ultimately, it’s really just a matter of what we choose: Ourselves, or AI?
