July 31st, 2026
By Keith Robinson
If the machine had the idea, is the scholarship yours?
That question ran through “The Human Factor: Legal Scholarship in an Algorithmic Age,” a discussion group at SEALS on July 31 moderated by Krista Bordatto. While the group never settled on how to identify authentic scholarly contribution, the discussion was lively and thought-provoking.
One speaker framed the stakes in terms IP lawyers will recognize. There is a sweat-of-the-brow argument buried in what we mean by scholarship. An author mines hundreds of sources and wrestles with ideas, and that effort has always been part of the claim to authorship. AI now does some of that work, and sometimes all of it. We need to arrive at a better understanding of what it means to do scholarship.
Disclosure drew the most attention. Participants wanted to know both whether and how scholars and students use these tools, largely because AI output sounds authoritative whether or not it is correct. Some legal writing programs ask students to certify they used no AI at all, though it is now close to impossible to research without touching an AI tool. One speaker noted that a judge in North Texas has written disclosure language that professors can borrow for their classrooms.
Publishing is shifting too. The Bluebook has a rule for citing AI, and law reviews are writing AI policies into author agreements. With more articles chasing fewer slots, one speaker asked whether editors will come to prefer authors who claim they do not use AI.
On assessment, the premise was that student work product is no longer reliable evidence of learning, so a student’s grasp of the process has to carry more weight. The methods raised included oral defenses, asking students to talk through the paper, grading metacognition, five-minute presentations, speed-dating rounds where students explain their work to each other, and annotating cases by hand.
Ultimately, scholarship is a form of self-expression, so it should matter to the author. Good scholarship should generate and share knowledge. It should also be a reflection of the author. That may be the most durable test we have.