Banning AI in Law School: We've Seen This Before

Historical Parallels: Cycles of Technological Anxiety

• In the 1920s, Harvard Law School banned typewriters, claiming they undermined discipline in legal drafting;
• In the 1980s, multiple law schools prohibited calculators, citing erosion of mental arithmetic and logical reasoning;
• In the early 2000s, University of Chicago Law School and others banned laptops in exams, fearing distraction and plagiarism.

None delayed adoption; all highlighted misalignment between pedagogical assessment and evolving tooling.

Current Bans: Implementation and Limitations

• NYU School of Law, Georgetown Law, and others have issued formal policies prohibiting LLM use (including GPT-4, Claude 3, Gemini 2.0) in assignments, exams, and classroom discussions;
• Detection relies on tools like Turnitin AI Detector—known for high false-positive rates (arXiv:2310.12977 reports 32% false detection on non-native English text);
• Faculty lack structured training in LLM-integrated pedagogy, turning bans into administrative shortcuts rather than capacity-building strategies.

Core Failure: Unupdated Assessment Paradigms

• Legal education’s core objectives—problem identification, argument construction, ethical reasoning—are orthogonal to text-generation speed;
• Bans conflate AI-assisted output with student capability, failing to distinguish ‘tool-mediated performance’ from ‘capability erosion’;
• The real risk lies not in AI itself, but in applying legacy standardized assessments (e.g., MPRE, Bar Exam formats) to cognition augmented by LLMs.