The ownership disclosure

Last week, the most prominent news in the higher education sector in the US was that the provost (and the chief academic officer) at Dartmouth College resigned. Dartmouth College is a premier institution of higher learning in the United States and while many readers of this newspaper may not really care about a higher education official of a US institution leaving his post, I believe that the reasons for his departure are relevant beyond the US and worthy of our reflection.

Dr Santiago Schnell, who served as the provost at Dartmouth for just over a year, was a well-regarded scholar in the field of mathematical biology. He had trained at Oxford and prior to his position at Dartmouth, he had worked at Indiana University, University of Michigan and University of Notre Dame. In August and September of this year, two news outlets – Semafor in August and The Dartmouth, the university student newspaper, in September – revealed that Dr Schnell had widely used generative AI in his writing without disclosing its use, including in journals and newspapers where there are clearly laid-out requirements for disclosure, or any use of generative AI is prohibited. When confronted, Dr Schnell acknowledged using AI for 'language refinement' or 'copyediting' but did not elaborate any further. As a result of the detailed story published in The Dartmouth student newspaper, many students and faculty at Dartmouth College were upset at what they saw was the blatant disregard for the university policies. They argued that if the student did such a thing (i.e. use AI and not disclose the use), they would get into serious trouble – so why is it that the chief academic officer can get away with it? There were others who found the entire episode distasteful – and antithetical to creativity, originality and transparency. The movement picked up steam and the university president tasked a faculty-led committee – comprising computer science faculty and humanities faculty – to further look into the matter.

The committee report came in the last week of October and it was damning. Dr Schnell had used generative AI in many of his recent publications, had violated policies of journals and was not forthcoming about those, and did not cooperate with the committee. Soon after the report was sent to the university president, Dr Schnell resigned and left the university.

There are many things to unpack here. First, the remarkable role of student journalists who are determined to hold university leadership to account. They deserve a lot of credit for their work and their commitment. Second, what incensed many people was that there should be same set of rules for everyone, and the faculty should be held to the same standards as they expect the students to abide by. But there is a bigger and perhaps nagging question about the pull of AI. Dr Schnell has had a distinguished career spanning well over two decades when he wrote outstanding papers and made important discoveries. His papers shaped ideas and created new knowledge. Why does someone like that choose to rely on something that takes away creativity and originality? Why does he need to rely on AI for language refinement when he did not need it for twenty something years? What pull does it have? And if someone does believe there is nothing wrong in using AI for 'language refinement' (or more) then why not declare it? Why not be open about it, or perhaps even be proud of it?

These are complex questions and perhaps we will never know the answers. But it is worth reflecting on how the notions of efficiency can force us to compromise the very foundations research, scholarship and inquiry. We should ask on our campuses and beyond: when we choose efficiency, are we OK with giving up creativity, originality and transparency? And should we be comfortable in doing so? Is that a good thing for the enterprise of learning, and society at large? And if not, then what guardrails should we create immediately?