A dirty little secret in higher education is that professors are not trained in how to teach. There are virtually no industries that hire a primary workforce without clear, and often degree-based, training in their most significant work function (except perhaps politicians!). Why this is the case largely has to do with a historical shift in professorial roles from teaching to research. Not coincidentally, evaluating what makes research and teaching significant is determined by the same people who conduct the research and teaching (i.e., professors are charged with evaluating and retaining themselves). It is no wonder, then, that teaching responsibilities and expectations are limited, with minimal oversight. The introduction of AI professors is bound to change things.
AI professors, in many ways, will be the best versions of the best professors students can have. AI professors will be realistic avatars that go far beyond the simple tutor model based on large language models, and will likely be here before anyone sees it coming. AI professors will: be available 24 hours, 7 days a week; have an exceedingly large bank of knowledge and experience that they can draw from to illustrate concepts; be complex responders to students’ learning styles and neurodivergence thereby providing truly personalized education with evidenced-based effective pedagogy; have the ability to assess and bring students along on any topic about which students desire to learn, thereby increasing access; teach content areas as well as durable skills such as critical thinking; and have updates in real time that fit the expectations and needs of the current workforce. A reasonable concern that has been raised is how to prevent AI professors from hallucinating or providing inaccurate information. One mechanism to guard against this is to ensure that the course and teaching that occur are within a closed system of content and have oversight by human professors. At the same time, it should be acknowledged that human professors are not immune to hallucinating or making up answers to questions. They just do it without oversight.
Human professors, on the other hand, will be able to serve students in a more personalized and interpersonal manner than is currently available. Instead of one to three-hour lectures a week for a class with upwards of hundreds of students (hardly personal), they will meet with smaller groups of students (say 20) multiple times a week, where they will get to know and develop meaningful relationships with many students (once university structures permit this). The responsibilities will include building relationships with students, facilitating the building of community among students, helping students network with other students and people in industries and careers in which students are interested. Another key function will be to teach and assist students with using AI to answer research and work-related questions in a multidisciplinary fashion. Presently, most college courses exist within disciplinary silos, which means that the most pressing questions are incompletely answered. For example, how to counter problematic climate change can likely be answered only via multiple disciplines (e.g., environmental science, political science, psychology, sociology, biology, etc.) rather than via an overly simplified singular approach. AI professors will be able to teach the multidisciplinary content, and human professors will be able to more readily bring the humanity that answers the ultimate why question (why it’s important to address the question to begin with, based on the personal and collective experiences of students). Of course, all of these new roles will come with expectations regarding professor training beyond content of their discipline, along with evaluation of their performance based on student success. Their onsite expectations will go from a few days a week for selective hours, with no teaching responsibilities in the summer, to five days a week for full days, year-round, with classroom and outside-the-classroom experiences with students. They will become the best coaches and teachers for students and play arguably the most critical role in student development.
Similar to AI convergence in the medical profession, the marriage between AI professors and human professors, a type of centaur model, will likely be a bumpy one. It will require the courage to look upon the current state of academia with the scientific prowess used to advance research, that is, objectively, using evidence to support contentions (e.g., link effective teaching with student learning). It will require professors to consider academic freedom as consisting of rights for professors and students, not primarily faculty. Historically, universities have not been bastions of meaningful, rapid change, particularly among the professorate. However, it has always been the case that the best professors responded to updates in technology in ways that integrated the new technology with their meaningful and effective manner of interacting with students. Those with an innovative mindset with student success at the center of outcome expectations tend to be the professors that students look back upon fondly. Universities that embrace this emerging approach will more likely succeed in the long-run and those that don’t are likely headed toward demise. The quicker universities understand it’s a “when” question rather than an “if” question about AI professors, the better chance they have to exist.