Should courts use artificial intelligence? Daniel Ho, the William Benjamin Scott and Luna M. Scott Professor of Law and Director of RegLab, and Olivia Martin, JD/PhD candidate, RegLab Law Student Fellow and Knight-Hennessy Scholar, tackled this question at the D.C. Circuit Judicial Conference with ...striking findings.

Their research shows frontier AI models can analyze thousands of regulatory comments in minutes and conduct comprehensive statutory surveys—tasks that traditionally take days. But these same models exhibit critical flaws, including hallucinations and accepting fabricated legal premises.

The stakes: Without hands-on experience, courts risk falling behind litigants already using these tools—creating a dangerous information asymmetry in the courtroom.

Read their full analysis: https://brnw.ch/21x5k7l

Research from the Deborah L. Rhode Center on the Legal Profession gets a close look in the most recent New York Times Docket newsletter.

The study asks: If court records are presumptively public, why are so many federal civil filings sealed? The answer points to a system in which sealing ...requests are often uncontested, leaving judges as the public’s principal safeguard.

As study co-author Professor Nora Freeman Engstrom told the NYT’s Adam Liptak, there are legitimate reasons for secrecy and real trade-offs. But judges need to do more than simply rubber-stamp requests to seal materials that shape their decisions.

Read more in the NYT (subscription may be required): https://brnw.ch/21x5hM3

Read the Rhode Center’s policy brief: https://brnw.ch/21x5hM4