A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by the Education Technology Insights Advisory Board.

Minerva Project

Teaching in the Age of AI

Art Markman

AI chatbots hit the national consciousness in 2022 and immediately discussions about the future of higher education began. The first wave of discussions was fear about academic misconduct. Even with the early commercially available chatbots, it was clear that they could do a credible job writing essays on a variety of topics, and that the technology would only improve. Over time, as the models improved, many of the kinds of assignments we give to students were ones that could easily be completed with the help of AI bots and agents.

At the same time, there were discussions about the role that AI was going to play in the world of work. As a result, in addition to worrying about academic misconduct, schools began to focus on the development of AI literacy skills. Early on, that meant teaching students how to write prompts. Now, there is also a focus on how to use a variety of AI agents to make work more productive.

Looking forward, higher education must transition from reacting to AI and teaching about AI to thinking about how to teach in a way that (1) diminishes the temptation to use AI inappropriately, (2) teaches for an uncertain future and (3) stops teaching about AI and starts teaching with AI.

Reducing the Motivation to Cheat

There are two fundamental reasons why students in college cheat. The first is that they think grades matter. Students have been conditioned since grade school to get the best grades they can. They have been told that grades influence important things like where they will be able to go to college. Even though almost nobody will ever look at a college student’s transcript and care about the grades they got, college students also care about getting good grades and so they cheat to improve their GPA.

Even students who think that grades matter typically don’t cheat when they are faced with assignments in their area of study that are helping them to learn skills they believe they will need after graduation. Unfortunately, the purpose of many assignments is not clear, and so students have trouble seeing how doing that work will teach them a skill that will ultimately be important for them to know.

A big reason why students feel that grades matter is that the only record they have of going to college is a transcript, which lists courses and grades. To de-emphasize grades, universities need to shift to giving students a record of the skills they are learning in their classes along with evidence that they have learned those skills. That evidence requires that faculty think carefully about the design of assignments and then communicate to students how those assignments and exams are helping students practice skills and giving them feedback on the current state of those skills. As I often tell faculty, “If your job is to shoot free throws in front of 19,000 people, you don’t pay someone else to do your free-throw shooting practice, but we never explain to students how the work we ask them to do is the equivalent of practicing free throws.”

Preparing Students for the Future

With the rapid development of AI models, we do not know what the future of work will look like. Already, some entry-level tasks that were given to recent college graduates are now things that can be automated with AI. In the future, even more complex tasks are likely to be automated.

To stay relevant in the workplace of the future, students need to hone their durable human-centered skills. Students need to be systematic and deep thinkers who communicate effectively, develop great ideas, think creatively, engage with civil society and develop a firm ethical foundation. To support this development, universities must identify the components of these crucial skills, identify where they are being taught across the curriculum and give students evidence of their growing competence.

“The future of education depends on the capacity of faculty to develop agents to support the teaching of their disciplines.”

The future belongs to those individuals who will best know how to navigate a rapidly changing world.

Teaching with AI

Faculty who supervises PhD students know that the best way to help a student gain proficiency in the complex skills that are central to being a scholar is to work with them directly as a mentor. Faculty advisors help their students to frame new problems. They check in with them frequently during the research process helping them to develop and hone the skills of their discipline and helping them to read the scholarly literature with a curious but skeptical eye.

Unfortunately, we cannot offer that individualized experience with a faculty member to every student. There are far too many students and far too few experts. AI offers a way to address that bottleneck. While AI is not yet up to the task of serving as a general mentor, it is possible to create agents for teaching specific knowledge and skills.

The future of education depends on the capacity of faculty to develop agents to support the teaching of their disciplines. Universities should develop templates to enable them to take particular lessons and turn them into activities for students that allow them to get specific tailored feedback. Rather than waiting for software companies to develop more generic applications, faculty should engage in the creation of systems that are directly related both to the specific knowledge and skills of their classes and that evaluate students with respect to the core skills their classes aim to teach. 

The articles from these contributors are based on their personal expertise and viewpoints, and do not necessarily reflect the opinions of their employers or affiliated organizations.

Weekly Brief