The University of Texas at Austin School of Law has issued a new eight-page policy memo directing faculty to prevent "cognitive deskilling" - the erosion of reasoning abilities - as students increasingly turn to artificial intelligence tools. The shift, outlined by Dean Robert "Bobby" Chesney last week, comes as law schools across the country grapple with how to integrate AI without undermining the foundational skills that define competent legal practice.
Chesney's memo does not reject AI outright. Instead, it calls for a deliberate balance: students must learn to use AI effectively, but only after they have done the hard analytical work themselves. "The lawyers most likely to flourish in the years ahead are those who both are adept at making wise use of AI capabilities and possess the same expert judgment and perspective that have always been the hallmarks of the best lawyers," he wrote.
The classroom, Chesney argued, remains the only environment where professors can be certain students are learning without AI assistance. His policy targets three specific areas: identifying which AI skills to teach, preserving the integrity of graded assessments, and stopping students from substituting AI-generated answers for the essential work of reading legal materials and building reasoning skills.
Socratic questioning as a safeguard
Chesney urged faculty to lean heavily on Socratic questioning - a method that forces students to articulate and defend their reasoning in real time. The technique makes it harder for students to rely on AI shortcuts without revealing gaps in their understanding. While AI can serve useful classroom functions, such as generating practice exams or acting as a tutor, Chesney warned that "AI also can be used as a shortcut that quickly and easily puts an answer into a student's hands, replacing deeper learning with a more-superficial semblance."
He acknowledged that students have always looked for shortcuts. But the temptation AI presents is different in scale. "AI will likely prove to be considerably more tempting than the traditional bypasses," he added. The memo frames this not as a technology problem to be policed, but as a pedagogical challenge that requires redesigning how law professors teach and assess.
A growing divide among law schools
The Texas approach contrasts sharply with other institutions. Last month, the University of California at Berkeley School of Law announced a blanket ban on AI use for class assignments and exams. That hard-line stance reflects a different calculation about the risks of AI in legal education - one that prioritizes preserving traditional skill-building over preparing students for an AI-augmented workplace.
Chesney's memo stakes out a middle ground. It accepts that AI for Legal work is now part of the profession's future, but insists that the core intellectual discipline of legal reasoning cannot be outsourced to machines during training. The policy pushes responsibility onto individual professors to design courses that make AI use appropriate in some contexts and impossible - or irrelevant - in others.
Why this matters for legal professionals
The Texas memo signals that law firms hiring in the coming years will encounter graduates trained under fundamentally different AI philosophies. Some will arrive with deep experience using AI tools under faculty supervision. Others will have been barred from using them at all. For hiring partners and practice group leaders, this means vetting candidates not just for credentials, but for how they were taught to think with - and without - AI. The split also raises questions about whether bar examiners and continuing legal education providers will need to address AI competency standards as the technology becomes embedded in daily legal work.
Your membership also unlocks:
