Different approaches to AI and faculty learning

Over the past three years we have been part of a variety of AI course and program redesign efforts. Along the way, we’ve learned some critical lessons about which formats are appropriate to solving specific problems. Different formats optimize for different priorities—time, scale, number of participants, and depth of engagement—and those priorities shape the kinds of outcomes that are realistically possible. This post is an attempt to make those trade-offs explicit. Rather than asking which format is “best” for AI work, we want to help practitioners match a format to a desired result.

Faculty Learning Community. Nik and his department chair Liahna Gordon have co-led a series of semester-long Faculty Learning Communities (FLCs) for other social scientists. This format will be familiar to many readers: a stable cohort that meets regularly over months. Research on faculty development consistently shows that FLCs are effective at producing sustained changes in teaching practice. In the context of AI, however, their most important affordance is time: time to work, revisit decisions, let ideas settle, and return having processed what felt unsettling only weeks earlier. One of our favorite early pieces on AI and teaching framed faculty responses through the stages of grief, an idea later explored on the Teaching in Higher Ed podcast. The metaphor resonates because meaningful pedagogical change involves loss—of familiar assignments, workflows, and assumptions. Grief processing takes time, and an FLC’s long duration allows faculty to encounter new ideas, sit with discomfort, and return ready to move forward. That same time horizon also supports trust. Trust grows slowly but can pay off: deeper trust enables more honest experimentation and peer feedback. The trade-off is scale. FLCs demand a significant time commitment, which limits participation. They are best suited for deep, sustained change with smaller groups, not rapid diffusion.

One Week Intensive. This is the format of the AI Retrofit which was conceived in this very blog and has since been implemented five times at Chico and adapted by multiple other campuses. It has also received external funding and been featured at conferences, partially because it has some of the best faculty ratings of any program we have ever worked on. We deal with grief in the intensive format, but we require people to speedrun it, which is not ideal. One strength of the intensive is scale. Meeting on zoom and having other staff to help in breakouts means we have regularly had cohorts of 50 with at least one other campus reaching 100. You can also run them consecutively meaning one well organized summer might reach 500+ faculty. The other strength of the intensive is allowing faculty to focus completely on a set of tasks without doing other parts of their jobs. This allows them to cohere together a set of related missions–something that is a bit harder in a FLC. For instance we have them map the AI disruptions to assignments and then move directly to solutions. We expose them to the power of AI to complete work which is often alarming, but immediately teach them how to leverage it for their own work which is a relief. Overall, this is a much more task focused format where faculty who are ready to work absolutely thrive, but it is not ideal for a faculty who seek a slower pace and a more intimate setting.  

Workshops and Speeches. Bluntly, these are not particularly effective and it has nothing to do with AI. For years we have known that a stand-alone workshop does not produce sustainable change and lessons learned are soon lost. However, we do believe there is a caveat with AI instruction. If you have an hour and most of that hour is spent having the people in attendance use an LLM–you can see a difference. This is essentially a replication of the most impactful part of our approach wherein you ask everyone in attendance to use a LLM to iterate and complete one of their assignments. This does not produce solutions, but it does help those in attendance realize the scale of the disruption. We have seen changes from this–especially for folks in academic leadership who do not have the same daily interaction with students. 

All of these approaches have a role to play as institutions chart paths forward with AI. The key decision is not which format is strongest in the abstract, but which priorities matter most to your organizational or individual goals and finding the best solution. Formats that optimize for in-depth time will sacrifice scale. Formats that optimize for scale will compress reflection. Formats that lower barriers to entry will select for readiness rather than depth. Ideally, institutions will find ways to support all approaches that meet the needs and requirements of units, programs, and individuals.