Richard Luedeman (University of Baltimore) reached out to me following my recent post on law faculty innovating with AI policies and shared this forthcoming article: “The AI-Accommodation Dilemma in Legal Education” (forthcoming 33 Michigan Technology Law Review (2027)). The abstract:
Two major trends are forcing American law schools to rethink their approach to learning assessments, and they are pulling in opposite directions. The first is the rise of generative artificial intelligence (“GAI”), which has made unsupervised assessments unreliable as measures of student learning and pushed faculty toward supervised, real-time alternatives (except, of course, where teaching GAI usage is itself a goal). The second is a dramatic increase in student requests for accommodations for academic disabilities—especially extended time and separate testing rooms—that make those supervised assessments costlier and more complicated to administer. Each trend, taken alone, would require significant institutional adaptation. Together, they have produced what this Article calls the AI-accommodation dilemma: the more that faculty embrace GAI-resistant assessments, the more accommodation obligations those assessments trigger; and the more burdensome those obligations become, the stronger the temptation to retreat to unsupervised formats that GAI has already compromised.
As law schools have begun publishing policies designed to insulate certain learning assessments from GAI, far too little has been said about the accompanying accommodation issues. To make progress on those issues, we must face them head on and resist simplistic for-or-against dichotomies. We should start from the open-minded assumption that well-intentioned faculty and administrators can hold a spectrum of nuanced views, and that many of them seek in good faith to help all students succeed and to balance the tensions between GAI concerns, accommodation obligations, and other important pedagogical goals such as frequent formative feedback. In addition, views understandably differ based on the types of courses (lecture, seminar, skills, or clinic), and thus assessment types, that particular faculty members are most concerned about. But the answer cannot be to unduly ration supervised assessments or to default to nearly automatic provision of one-size-fits-all accommodations. Instead, this Article argues that federal law grants law schools more flexibility to solve this challenge than they typically exercise in this domain. Employing that flexibility thoughtfully and creatively, with greater faculty involvement and case-by-case analysis, is not only legally permissible but, in the age of GAI, pedagogically necessary.
