Alongside calls for campus AI literacy instruction and critiques of AI use by college students, there is a quieter, but no less important, discussion centering on faculty AI familiarity and proficiency. The simple formula of most college learning structures is that instructors can impart knowledge, skills, and abilities to students in large part by sharing their own expertise. Generative AI’s sudden growth and consistent evolutions have upended the traditional model of instructor as expert and student as novice, at least in terms of generative AI. As one instructor in our recent national survey on AI skills put it, “I’m learning how to use [AI] along with the students.”

Acknowledging that some instructors reject AI outright and see no place for it in higher education, there are increasing numbers of faculty members seeking to better understand, utilize, and integrate AI into their own practice as teachers and researchers. While institutional centers for teaching and learning, libraries, and IT offices are all trying to help instructors adopt or adapt to AI, the broad impact of these efforts remains unknown.

To better understand where instructors look to for AI-related support, we asked the 500 participants in our national survey how frequently they use various resources when learning about generative AI or developing their own AI skills. We found that on average instructors most frequently go it alone when learning about AI. This independence in their learning aligns with instructors’ hesitancy to engage in faculty development more generally, often due to anxieties that doing so implies a weakness or professional deficiency.

For administrators who aim to promote faculty learning on AI, this finding may be discouraging. The data, however, do suggest that departmental approaches may be a key point of leverage. The instructors we surveyed more frequently leaned on departmental colleagues and inter-departmental learning communities as learning resources over services offered by administrative offices, though even these learning communities were still not often utilized. Institutions looking to advance faculty adoption and adaptation of AI should seek to channel efforts through these peer learning mechanisms to better align with instructors’ existing practices.

In our national survey on AI skills, we found instructors prioritize teaching students how to use AI responsibly above all other AI-related skills, although almost one half of those same instructors do not actively teach that skill in their courses. This dissonance may be an important place for faculty professional development around AI to begin because it leverages a high-priority skill in which there is a critical mass of peers already integrating that skill into their teaching. As instructors become more fluent in AI, there will be greater opportunity for the peer learning with and through faculty colleagues that instructors already use more frequently.

Figure 1: Instructors rely on themselves most frequently to learn about AI