Supporting Faculty in the Age of AI: Takeaways for Campus Leaders

Author: NCFDD

Faculty are working through questions about AI that have immediate consequences for their teaching and workload. An assignment may need to change so students can demonstrate their understanding. A tool that promises to save time may require considerable work to evaluate. For campus leaders, supporting those decisions means understanding what faculty are encountering in practice.

Those experiences were the focus of NCFDD’s recent webinar, What Faculty Are Telling Us About AI in Higher Education. Drawing on findings from Faculty Perspectives on AI in Higher Education, the discussion brought together faculty and higher education experts to examine how AI is affecting their work and where institutional support could help.

Moderated by Lisa Hanasono, PhD, NCFDD’s Academic Director of Training and Content, the conversation offered several takeaways for leaders considering their approach to supporting faculty with AI-related changes. Panelists shared specific examples while raising broader questions about what students need to learn and how institutions can give faculty time to respond.

Faculty support should strengthen informed judgment about AI

Before deciding whether to use AI for a particular task, faculty need to understand its limitations.

Geetanjali Datta, MPH, ScD, founder and principal researcher at Sagebrook Research, described several challenges that deserve attention. AI can produce false information or give different answers to the same question. It can also lean toward agreeing with a user, making its responses less useful when someone needs a rigorous critique. For researchers, privacy requirements and agreements governing data use introduce further considerations.

Datta emphasized that skepticism can be an appropriate response to these limitations.

“Academics are very well trained in critical appraisal, so it makes complete sense that faculty are approaching new tools with curiosity and skepticism, even after some use,” she said.

She offered a familiar scenario to help faculty evaluate possible uses: imagine a new, eager research or teaching assistant who has AI’s known shortcomings. Consider which tasks you would feel comfortable assigning that assistant and how you would review the results.

That exercise brings the decision back to faculty expertise. An instructor might find a tool useful for generating possible examples while deciding that another task requires their own sustained attention. A researcher may identify a potential application that cannot proceed because of restrictions on the information involved.

For institutions planning faculty development opportunities, these are useful questions to build into programming. Faculty need opportunities to examine applications in their own work and determine whether the benefits justify the risks and effort involved.

Assessment changes can make student reasoning more visible

Marlene Shaw-Gallagher, MS, PA-C, an associate professor and lead clinical coordinator in the Physician Assistant Program at the University of Detroit Mercy, described how increasingly polished student papers prompted her program to reconsider its assessments.

“We wanted to make sure that we were truly looking at our students’ experiences and hearing from them in their own voice what they were doing,” she explained.

The program replaced those written assignments with oral case presentations followed by questions. Students returning from clinical rotations can use AI to help prepare their slides, but they must explain their clinical reasoning during the discussion. Faculty and classmates ask how they reached a diagnosis or developed a treatment plan.

“Then there’s a Q&A session, where that’s where it becomes more visible, what their learning is,” Shaw-Gallagher said.

The example offers a starting point for reconsidering assessment in the age of AI: identify what students need to demonstrate, then examine whether the assignment gives faculty enough evidence of that learning. The appropriate format will depend on the course and its learning goals.

José Antonio Bowen, PhD, former president of Goucher College and co-author of Teaching with AI, raised a related concern. Students may use AI in ways that comply with course rules while bypassing work that helps them learn. Having a tool summarize lecture notes, he suggested, could remove some of the effort involved in making sense of the material.

His question for educators was broader: “How do we redesign learning?”

Working through that question takes disciplinary knowledge and opportunities to discuss what is happening in actual courses. Institutions can support the process by giving faculty time to examine assignments with colleagues and consider what evidence of understanding they need.

Time savings need to be considered alongside the learning curve

Shaw-Gallagher also shared an example of AI helping with a substantial administrative responsibility: coordinating approximately 60 students across 10 clinical rotations each year.

She described developing a scheduling tool that could account for placement availability and help adjust the schedule when a preceptor could no longer take a student. Building that process required experimentation, but she reported that it had saved her considerable time.

Her approach begins with the individual task.

“I really look at it as a task-by-task analysis of what can I use AI to help me, and where is it going to cause more work for me?” she said.

That question gives institutions a useful starting point for evaluating potential productivity gains. Developing a workable process takes time, and faculty must still evaluate whether the results meet their needs.

Bowen also pointed to the possibility that discovering more uses for AI can intensify work by creating a sense that there is always something else to try.

These experiences add context to conversations about AI and faculty workload. A useful application in one workflow does not establish what will happen to a faculty member’s overall workload. Leaders can make room for evaluation by asking what a proposed change requires and whether it actually addresses a problem faculty want to solve.

Human connection deserves deliberate attention

Toward the end of the conversation, Bowen returned to the relationships between faculty and students. He encouraged institutions to consider how they could strengthen the human interactions they value.

“You can’t talk to every student in a 500-student class every day, but 500 students can watch you talk to one student,” he said.

His examples were practical: arrive a few minutes early and make yourself available for conversation. Being visibly interested in students can help communicate that approaching an instructor is welcome, even when speaking with everyone individually is impossible.

Bowen connected this attention to relationships with institutional decisions about workload. If a tool creates additional time, how will that time be used?

Making education and human relationships better, he argued, “is not going to happen automatically. We need to invest in that.”

For provosts and faculty development leaders, that raises a concrete question: how can support for faculty help them devote attention to the parts of teaching that their institution values most? The answer should inform decisions about professional development and the time faculty have to put new approaches into practice.

Watch the full webinar, What Faculty Are Telling Us About AI in Higher Education, for the panel discussion and audience Q&A.

Faculty_Perspectives_AI_Higher Education_Report