As co-author of a book called Getting to Maybe, it’s no surprise that my answer to this question is that it is and it isn’t. But here’s some actual evidence on the question produced in a study using University of Minnesota law students by four faculty members at the University of Minnesota Law School (listed below). As described in their paper, Artificial Intelligence and Human Reasoning, which you can find here, they reached some interesting conclusions about how working with AI at initial stages of a project didn’t interfere with later tests of comprehension of the material. Yet they also observed that bringing AI in early tended to improve low quality first drafts while dragging down first rate ones. Much to ponder.

Bednar, Nicholas and Cleveland, David R. and Erbsen, Allan and Schwarcz, Daniel, Artificial Intelligence and Human Legal Reasoning (April 05, 2026). Minnesota Legal Studies Research Paper 2026-21, Available at SSRN: https://ssrn.com/abstract=6525800 or http://dx.doi.org/10.2139/ssrn.6525800