Justice Sotomayor Warns Future Lawyers: Master AI, Heed the Ethics
Justice Sonia Sotomayor told law students to learn AI thoroughly and warned of mounting ethical pitfalls.
Why it matters: AI adoption in the legal field is accelerating, raising questions of professional responsibility and risk. Legal professionals must understand both how to use and how to scrutinize AI tools to meet emerging standards and avoid costly mistakes.
- Justice Sotomayor urged students not to graduate without learning AI as a key legal tool.
- She cited AI's 70%+ accuracy in predicting Supreme Court rulings over nearly 200 years.
- Federal judges are increasingly using AI for case prep and draft opinions, per multiple 2026 reports.
- Over 1,200 global sanctions—800 in the U.S.—have involved errors in AI-generated court filings.
U.S. Supreme Court Justice Sonia Sotomayor addressed law students in April 2026, calling for rigorous education in artificial intelligence and sober attention to legal ethics. Her remarks come as legal technology reshapes daily practice—and as the judiciary confronts new uncertainties.
- “Do not graduate this institution without learning how to master AI as a tool,” Sotomayor told students, emphasizing both skill and a sense of duty when deploying new technology (Washington Examiner, April 6, 2026).
- Sotomayor referenced research showing that AI models can predict Supreme Court outcomes with more than 70% accuracy for cases from 1816 to 2015—a statistic she framed as both impressive and cause for professional reflection.
- Recent reporting indicates a marked increase in AI use among federal judges, including for preparing cases and drafting preliminary opinions, as noted in surveys and public statements in 2026 (Washington Examiner).
- Sotomayor likened AI capabilities to those of a "sophisticated human," warning that the technology can replicate both the strengths and weaknesses of its users—including the risk of propagating bias or error.
- Legal organizations are seeing a spike in sanctions tied to faulty AI-generated submissions—over 1,200 worldwide, including 800 in the U.S. by April 2026 (WUSF).
Sotomayor’s speech reflects a judicial consensus: as AI becomes routine in courts and law firms, lawyers are expected to balance innovation with caution and uphold professional standards in an evolving landscape.
By the numbers:
- 70% — AI's historical accuracy in predicting Supreme Court outcomes (1816-2015)
- 1,200+ — Sanctions worldwide for AI-related errors as of April 2026
- 800 — U.S. sanctions reported for AI-generated court filing mistakes
Yes, but: Some cited reports come from less prominent sources; broader confirmation from mainstream legal outlets would strengthen the evidence base.
What's next: Expect more law schools and bar associations to update curricula with AI literacy and ethics training in response to judicial guidance.