EDITORIAL: U.Va. needs to promote better artificial intelligence policies

You are effectively an anomaly if you did not use artificial intelligence over this past finals season — for better or for worse, almost every college student does in some capacity. As it has invaded so much of our lives, it is clear that AI has the potential to further revolutionize workplaces, the government and schools. Some believe it can even create art. Nevertheless, faculty have difficulty agreeing if AI use is detrimental to critical thinking or actually enhances it, leading to a wide variety of classroom policies. While some students vote on what their classroom policy is, some only receive a note at the bottom of their syllabus forbidding AI use.

But however faculty see this new technology, the problem arises when professors are forced to create their own AI policies without any substantial guidance or structure.  The AI Task Force was meant to help solve this issue, but as it understandably did not set out to create any real policy, it fundamentally falls short of providing real change. The University needs to do more to lay out a standardized, departmental process through which professors and students can discursively formulate their own policies and communicate them to students clearly.

The University AI Task Force has been the University’s primary answer to the question of AI use in the classroom. While the McIntire School of Commerce and the School of Data Science both have similar agencies, they clearly draw from the Provost’s Task Force. A quick look at the Task Force’s website will reveal no actionable solutions. The one and only report provided by this agency is from summer 2023. While the information includes valuable surveys taken of students and faculty on their AI use tendencies, everything must be caveated by the fact that the Task Force published it two years ago. Unless you have been living under a rock, you will have probably experienced the recent rapid growth in the power of AI. It is no understatement to say that while policies from two years ago may provide a valid springboard for further development, they are not at all sufficient to be applied to our University classrooms today. 

The shortcomings of the Task Force create an opaque situation where students suffer. One year after the inception of the Task Force, many students at this University feel like they have not been included in the implementation, or conversely, eradication, of AI in the classroom. One year later, many students use AI on their assignments. To be sure, being caught definitively using AI to violate the honor code is a troublesome cheating offense. But without using AI, students know that they are set at a disadvantage to their classmates who likely use the technology. Thus, students are caught in a terrible limbo — to cheat or not to cheat. This is a failure of policy. 

To faculty, the Task Force falls short because it claims to give professors total autonomy over their AI policies, but by leaving this process completely decentralized, faculty are left without all the necessary information and tools to make full use of AI in their classes. They are similarly left on a fence, trying to decipher what the Task Force exactly recommends for their unique course and how the content might benefit from AI. Departmentalized policies would allow for flexibility by subject, yet the task force does not give faculty the toolkit to properly reframe their classroom policies to accommodate AI. 

The Task Force, however, is not without merit. The “Guidance for Faculty and Students” section of the Task Force website lays out the general idea of how AI should be implemented in classrooms and does a good job promoting AI policies beneficial to students and faculty. It describes how faculty should find ways to integrate AI into the classroom in a responsible manner and communicate their expectations clearly to students. Further, the website accurately states that students should be included in the formation of AI policies in order to generate a better relationship over the use of such technology. The guidelines given to professors do attempt to solve one of the most pressing issues in academia, but we as students have experienced AI policies which are often disparate and without cohesion. They are often not influenced by the subject material or department. Clearly, the right idea is here, yet formal integration into classrooms is sorely missing.

A better method of controlling AI use in classrooms comes with collaboration and communication — the University should set a clear policy-making process for professors. This new process should build on what the Task Force suggested — incorporating students, giving professors the tools to understand AI’s potential and flexing for the unique nature of courses — but actually have departments implement a standardized process. Due to different AI potential in different subjects, departments should formulate their own technological assistance policies so that students in each major know broadly what is expected, with exceptions as individual professors see fit. This could look like the English department using AI for finding literature for projects while barring it from writing and the Computer Science department using it for some busy-work like debugging. This policy would exist to find a better middle ground, facilitating both faculty discretion and student realities.

A defined, clear process for creating AI policy would allow faculty to maintain authority over their syllabi while creating a system that avoids any miscommunication. Students need not worry about how they will use AI, because they will have a hand in forging their guidelines. Approaching next semester, AI will only grow more powerful, and this institution of higher learning needs to generate new ways to grow with AI — rather than rely on data from two years ago. 

The Cavalier Daily Editorial Board is composed of the Executive Editor, the Editor-in-Chief, the two Opinion Editors, a Senior Associate and an Opinion Columnist. The board can be reached at eb@cavalierdaily.com.