The Faculty of Arts and Sciences, the largest academic division at Harvard, published its first set of guidelines for instructors on using generative AI in their classes. The Office of Undergraduate Education’s instructions are extensive and provide general knowledge on how generative AI operates and its potential academic uses.
A fully-encouraging policy, a maximally restrictive policy, and a mixed approach are the three methods that professors can use towards using AI in their courses, respectively. The guideline does not impose one AI policy across the FAS, instead, it offers specific terminology for these three options.
One grounding basis underlying the recommendations, according to Christopher W. Stubbs, dean of science, is that faculty have ownership over their courses. He said, ” I don’t think there is a one-size-fits-all course policy here. What we are requesting of the faculty is that they become informed, that they comprehend the impact this has on the learning objectives for their courses, and then, crucially, that they communicate to students clearly and frequently what their course policy is.”
The FAS guidelines also expand on university-wide AI policies that were released in July and put a strong emphasis on safeguarding private information. Faculty are not to enter student work into AI systems, according to FAS guidance. Stubbs pointed out that third-party AI platforms own both user-generated prompts and computer-generated responses.
Early last month, the school held two informational meetings for professors on the effects of generative AI in STEM and writing courses. The sessions cover potential uses of AI as an aid for learning, including real-time information synthesis, code generation, and argument evaluation.