Throughout this book, each project addresses unique aspects of teaching and learning, and highlights various opportunities and pitfalls in using AI tools in the workflows of college faculty. Yet we find commonalities across projects, which we delineate below. These insights were aggregated both from the resources that the authors created and follow-up group discussions held after the completion of their projects.

AI in Education Grants Facilitated Deeper Analysis of Learning Outcomes

It was noteworthy that many faculty indicated that they were already intending to carry out similar projects, but that their involvement in this program prompted them to critically examine how they would measure the success of their projects. Towards this goal, the program encouraged applicants to systematically document their use of AI and identify potential improvements that future instructors could learn from. That is, it shifted faculty from a Transactional perspective of simply using AI output to a Transformational perspective of engaging more deeply with that output (see Navigating Changing University Policies).

For example, in our follow-up conversation, an author noted, “We would eventually incorporate [this AI tool]… but being able to do it with the whole class, able to do the survey for student feedback; I wouldn’t have done that otherwise. So that was exciting to hear student perspectives.” Other authors echoed a similar sentiment. Many projects incorporated comparisons with more traditional non-AI methods to determine the added benefits of incorporating AI (e.g., AI-Assisted Scientific Literature Reviews, & AI-Assisted Article Review Assignment), and some focused on engaging with AI output in critical ways to better understand (or get their students to better understand) the guardrails in which AI could be used effectively (e.g., AI-Supported Coding Assignments and AI-Assisted Design of Polymerase Chain Reaction Primers).

Finding the Right AI Tool in a Rapidly Evolving Landscape

Many authors indicated the challenges of finding the right tool and using it effectively. For example, one author stated, “For most stuff I use enterprise edition of chatGPT. Copilot it just took more effort; it was just kind of a mess. [Now] I use otter.ai for making transcripts of files and summarizing. And I use Claude for brainstorming… I’m starting to think about what sorts of tools might be most helpful for certain applications.”

This was particularly true as we consider the constant evolution of these tools and the non-reproducibility of tools over time. Another author documented that “[the capabilities] varied even from lab section to lab section, so I think maybe there was an update midweek… It’s the limitation of something that’s always evolving, and what might work one semester might not work a different semester because of the specific ways you’re using it.” University policies that limit or encourage particular tools can make this even more challenging; some awardees described working through limitations with Microsoft Copilot even if they felt another tool could perform better. Therefore, it is particularly important to document the existing tools and have an evolving understanding of which tools are best suited for particular purposes.

Projects Were Catalysts for Productivity and Creativity with AI

It was clear that many faculty saw value in the productivity these AI tools could have in their teaching and research (e.g., AI-Assisted Exam Creation & Summarizing Student Feedback with AI). It was not surprising that most authors who attended the follow-up discussion expressed a desire to continue using AI or adapting their use in future terms. One author nicely summarized, “Having this grant allowed me to delve into this area that I was actually a little nervous about doing because I knew nothing about AI and how to use it… [this grant] allowed me to kind of sit back and think, Well, what am I struggling with and could it help? [Now] I’m using it all the time… it’s really kind of opened up my world in terms of how to be more efficient, how to be more thoughtful about work I’m doing in the classroom.” Many discussed how trivial, unenjoyable tasks could be done more efficiently with AI. For example, another author noted, “My syllabus dates are always something I mess up… but [with AI] it was all fine with no errors… It’s just those little things…” Alongside these examples of using AI to increase productivity, many of our authors predict that being able to use AI for rote tasks will accentuate creativity as an important skill in educators’ lives and in our students’ future success.

Students and Instructors Have Concerns About AI Use

While these faculty believed that instructors and students should engage more with AI tools, there was also careful consideration of students’ and colleagues concerns in using AI. It was clear that many students were concerned with using AI, given the predominant message that AI use was inherently cheating. This perception changed as students were encouraged to use AI tools as part of the expectations of the courses. As one project nicely documented (AI-Assisted Article Review Assignment) students said things like: “I was excited (about using AI in an assignment). I did not know I can do that so this was a pretty eye-opening experience for me” and “This course and assignment has demonstrated how AI should be used in addition to creating my own work and not as a substitute.” To address students’ privacy concerns, another project (AI-Facilitated Job Application Materials) provided students options to still engage in AI-assisted professional development, but provide generic information if they were concerned about entering their personal information and perspectives into the tool. Overall, understanding the perspectives of educators and students when implementing AI tools is particularly important given the novelty and nature of these tools.

We’re Only Scratching the Surface of Educational Innovation in AI

Nearly all of the project applications were unique, which was somewhat unexpected. As we organized these grants, we envisioned a few use cases that instructors would be attracted to within a similar subset of potential ideas. But we were astounded by the different ways people attempted to use these AI tools across only 16 projects. Similarly, an AI in Education guidebook, produced by our University and Grammarly, included lesson and activity ideas that have only minor overlap with this resource. Thus, it seems there is much more to be done and discovered. So to our readers, we will simply say that we look forward to seeing what you come up with next.