BY D.J. HOPKINS AND ZACH JUSTUS

It’s time for your department to produce its annual assessment report. You and your colleagues have agreed on the learning outcomes to measure, and you have the data, but no one looks forward to this task, so you put it off. You answer emails, do some compliance training, and go to get another coffee. En route, you run into a friend from a different unit tasked with the same report. You expect to commiserate over the drudgery of the task, but instead you find they, or rather they in coordination with Codex running on their computer, have already compiled the essays, analyzed them, put the results into the form, and submitted it. It took them thirty minutes, while you have the rest of the day blocked just to get started.

This hypothetical scenario is playing out across every domain of the academy, and this dual track of faculty work is what inspired our title. You already work at two universities.

From the beginning of AI’s disruption of higher education, the focus has been on academic integrity. Meanwhile, we’ve missed a revolution in faculty effort resulting in another rupture, in addition to the familiar divides between science and the humanities and faculty in tenure-track and contingent faculty appointments.

We’re talking about AI users versus AI refusers, or colleagues who simply work as they always have, without making a conscious choice about AI use. How is this new divide surfacing in the academic workplace? In a word, productivity, distributed across every domain of faculty effort.

Research. AI can be used to locate grant opportunities, produce the prose of funding applications and tables to support it, and even fill out required forms. Combined with the experience and expertise of a professor, these systems can allow scholars to dramatically increase their research output. While the AAUP has introduced a long list of questions that researchers should ask themselves about using AI in their work, how many faculty members do you know who are familiar with this insightful list of AI cautions? And how many AI power-users would be willing to tap the brake on their productivity when the benefits seem obvious and the consequences are ill defined?

Teaching. AI is being used by many instructors to create slide decks, essay prompts, rubrics, and other course collateral. We all know some who have started to outsource part (perhaps all) of their grading to AI. And still others have spun up AI course tutors that serve as a first line of support for students. Such AI assistants can compound a research advantage, freeing up faculty time. While numerous sources document the value of in-person teaching and mentoring, faculty members who invest significant time in classroom community-building have been losing the research race to colleagues who don’t for years. Of course, many faculty members try to split the difference, but we expect that this productivity gap will only widen.

Service. The scenario outlined in our introduction is not science fiction. It is a reality we are living now. “Assessment” has been a four-letter word among faculty for years, so the temptation to use AI to expedite what is already viewed by many as drudgery is clear. New AI tools make it easy to analyze large datasets like teaching evaluations and statistics on degree-level learning outcomes. After an AI app generates a spreadsheet, some faculty members will be eager to evaluate that data and write up a conclusion; some will simply prompt the AI to compile the report for them. And there are those who will print out two hundred pages of student success data, clip them into a three-ring binder, then sit down to read. We know who’ll finish first.

All these efficiency advantages of AI use compound around the single concept of productivity. What are the implications when each university now exists in two productivity realities?

For one thing, faculty review for tenure and promotion is about to get even more complicated. Should committees evaluate a faculty member who has used AI for all their work (and published twice as much as department criteria expect) by the same set of standards as an AI conscientious objector? Or, more likely, a colleague who simply has not adopted AI tools in their workflow? Alternatively, will faculty encounter tenure committees with members for whom any whiff of AI usage—even if only to support workflow, not to produce text—is evidence that an entire portfolio of academic work is tainted and should be thrown out? These are new questions we have not had to consider before.

Each of us has different opinions on this topic, but which side of this divide you identify with doesn’t matter. The point is that the fault line has formed. The question becomes, What can we do about the widening chasm?

One way to address this emerging divide is by introducing new criteria to the tenure and promotion policies at our universities. How do we prevent Professor Curmudgeon from arguing that a publication that discloses AI use should be invalidated and encouraging the tenure committee to vote against a candidate? Or Professor Antagonist from criticizing a candidate for averaging only two articles a year when he and other AI users now average five? And should we agree to set limits on productivity expectations? If the number of publications starts to creep up in one field because the emerging norm is to use AI, then publication expectations may start to creep up for everyone who wants to keep up, regardless of discipline or AI use.

We’re not here to tell you whether or how to integrate AI into your research, we’re letting you know that, in case you missed it, you already work at a university where some of your colleagues are doing everything they can with AI. It is within our power as faculty to set parameters for AI use on our campuses, including how AI use is evaluated in the tenure and promotion process. We can collectively decide how our universities will navigate this change. But shared governance processes are slow. Now is the time to start thinking about AI and productivity before the divide becomes too vast to easily bridge.

D.J. Hopkins is a professor at San Diego State University, where he is the director of AI programming for the Center for Teaching and Learning. He has led a faculty development portfolio focused on AI adaptation, including workshops, learning communities, online resources, and a podcast. His scholarship frames generative AI in the discourses of performance studies and digital humanities. He regularly speaks on AI literacy in higher ed.

Zach Justus is the director of faculty development, professor of communication arts and sciences at California State University, Chico, and the CSU AI Academic Innovation Fellow for the CSU system. On the topics of AI and higher ed, Zach has collaborated to produce Inside Higher Ed and EdSource articles, several webinars, an ongoing blog/podcast/newsletter, and a series of conference and keynote presentations.