Michel Foucault famously cited the concept of a panopticon to describe a system of power that relies not on constant surveillance, but the possibility of being watched. Each prisoner doesn’t need to be watched at every hour; rather, uncertainty is the enforcement mechanism. With the accessibility of generative artificial intelligence, many universities have attempted a similar structure — while no clear enforcement method is outlined, the novelty and uncertainty around the new technology is enough to warrant caution.
However, as students and professors alike have begun to recognize the inability to accurately distinguish between organic student work and generative AI, this panopticon is slowly disintegrating. AI detector models have elicited fluctuating anxiety, and ethical conflicts arise among students as behavior that is praised in one classroom might be reprimanded in another.
Professors need support in their pursuit of classroom policies around generative AI that are not only logical to the class but also enforceable. Questions of the role of generative AI in higher education have become increasingly pressing, and professors are expected to take on the place of architect and guard, deciding both the role of AI content and how it would be maintained. UT needs to better equip professors to build these policies and identify ways to enforce them, through the consolidation of resources and platforms for discussion.
“I want to feel like my university supports my decision to not use AI in the classroom and to find ways to ensure that the students are getting the best education that I can give them,” English professor Deb Olin Unferth said. “I feel like there’s been a lot of talk about how great AI is, and if we can have a little bit of conversation, or a little bit of support for the other direction, there are things that you can learn without AI.”
Some universities feel it remains acceptable to use generative AI for supplemental purposes. However, humanities departments may particularly be suffering from this uncertainty due to the commonality of long-form, take-home assignments like essays or short stories. Professors of these writing-heavy courses may have a harder time outlining an entirely no-use policy around generative AI because there has yet to be a particular way to enforce this. Philosophy professor Kathleen Higgins discusses the difficulty of identifying uses of generative AI in student work.
“Obviously you can’t establish from the fact that somebody uses a lot of dashes in their writing that AI helped, even though apparently AI has more of an enthusiasm for them than the average student does. It’s impossible to prove that somebody has used AI,” Higgins said.
Now, it may be a strong call to expect the UT administration to resolve an issue that doesn’t have an answer yet. There is no solution to the current blackbox nature of AI, and a study of over 12 of the most common AI checker programs have found them neither reliable nor accurate. There also most likely isn’t a cookie-cutter solution that can be applied across departments and fields of study, at least in the foreseeable future.
“Anything that’s too rigid or designed for universal application, I don’t think is very helpful,” Higgins said. “There is a certain amount of creativity in the use of AI that people can engage with and … the problem is it blocking other kinds of displays of creativity.”
However, ignoring the issue will not bring about a more productive solution. Professors have found ways to pursue enforcement, a problem that remains pressing and still requires a solution in the present. Unferth has restructured writing assignments to have more in-class elements, as well as hand-written requirements. Unferth explains how she came about her no-use policy and changes for her classes.
“I have come up with it mostly on my own through conversations with other professors,” Unferth said. “I’ve been in constant conversation with other professors, not only at this university, but in many universities.”
Although there isn’t one clear-cut solution, UT can help by providing more support for professors who are facing this issue head-on. While each class is different and requires its own nuanced solution, UT should help consolidate resources and information to support the rather large decision that has ultimately burdened each individual professor.
Tanya Narwekar is a philosophy and economics senior from Coppell, Texas.
