‘I can simply copy-paste material, so do I truly need to learn?’

In what ways is generative AI influencing higher education? Faculty, students, and staff from various departments within Harvard convened recently to address this pressing inquiry.

Participants at the half-day event, co-sponsored by the Digital Data Design Institute at Harvard, the Office of the Vice Provost for Research,, and the Office of the Vice Provost for Advances in Learning, grappled with the ways generative AI is altering the competencies required for student success. The primary takeaway: AI necessitates a fundamental reassessment of higher education’s objectives, methodologies, and principles.

“This is an extraordinary chance for everyone at Harvard to envision how our working methods, teaching techniques, learning processes, and research approaches can be radically transformed,” expressed Karim Lakhani, Dorothy and Michael Hintze Professor of Business Administration and founding chair of the Digital Data Design Institute. “Our mission of veritas can indeed be enhanced through these tools.”

“We must understand that having access to information is not equivalent to learning, and it definitely does not equal active and sustained learning.”

Nonie K. Lesaux, dean of the faculty

Harvard University Provost and Dane Professor of Law John Manning characterized the gathering at Klarman Hall as a “One Harvard” moment.

The symposium brought to light differing opinions regarding the acceptable applications of AI in academia. Rebecca Nesson, the panel moderator and dean for academic programs at the John A. Paulson School of Engineering and Applied Sciences, surveyed the audience: Is it acceptable for students to utilize AI for summarizing class readings instead of doing the readings themselves? Is it appropriate for a professor to employ AI to draft a recommendation letter? What about conducting a preliminary assessment of student assignments? The audience was divided.

In a panel on generative AI and the future of education, the speakers contended with the implications of truly learning something in an AI-influenced environment.

“We must be clear that having access to information does not equate with learning, and it certainly is not comparable to active and sustained learning,” stated Nonie K. Lesaux, dean of the faculty and Roy E. Larsen Professor of Education and Human Development at Harvard Graduate School of Education.

Christopher W. Stubbs, Samuel C. Moncher Professor of Physics and of Astronomy at FAS and senior advisor on generative AI, noted that faculty at FAS exhibit varying levels of AI integration and policy acceptance in the classroom: Some tailor their teaching around it, while others prohibit it entirely. “We have a long journey ahead to help our colleagues understand what this entails, to fundamentally revisit our educational objectives, and then reconstruct our approach,” he remarked.

Iavor Bojinov, assistant professor and Richard Hodgson Fellow at Harvard Business School, instructs the inaugural AI-native course at the School, “Data Science and AI for Leaders.” Students leverage AI to prepare for classes, respond to inquiries, and develop their own enterprises. “People have a very favorable outlook on these tools, but it has also made cheating significantly easier. Numerous students approached me and asked, ‘I can just copy-paste material, so do I really have to learn this?’ This is an issue we are seriously contemplating,” he remarked.

Notwithstanding these unresolved questions, researchers throughout Harvard are utilizing generative AI to accomplish extraordinary breakthroughs. Alberto Cavallo, Thomas S. Murphy Professor of Business Administration and co-director of the Pricing Lab at the Digital Data Design Institute, employed it to determine the countries of origin for thousands of products to approximate the effects of tariffs. Rachel Carmody, Thomas D. Cabot Associate Professor of Human Evolutionary Biology at FAS, presented AI research that sifted through 40 terabytes of metagenomic data for an analysis of the potential extinction of specific gut bacteria in developed nations. “It’s a trillion-piece jigsaw puzzle,” she noted. “With generative AI, we can obtain results in a week, whereas it would have taken us a year just a year ago.”

The symposium represents the latest in Harvard’s initiatives to enhance teaching, learning, and research through the inventive application of AI. Harvard Business School established the Digital Data Design Institute in 2022 as a global research center, delivering data-driven insights on how AI is reshaping work and the economy. In 2023, Harvard University Information Technology, in collaboration with the Vice Provost for Advances in Learning, the Faculty of Arts and Sciences, and faculty and staff from across the University, initialized an AI “sandbox” allowing users to safely experiment with various large language models.

The symposium concluded with student presentations and a reception featuring interactive demonstrations of AI applications. FAS, the SEAS, HUIT, and the Harvard Library were also co-sponsors of the event.