Despite numerous mistrials resulting from artificial intelligence in the courtroom, we learn—courtesy of Super Punch—that Los Angeles county, the largest civil justice system in the US is running a pilot programme that allows judges to use an AI tool, called Learned Hand, to draft legal opines and tentative rulings, informed by precedent and the individual jurist’s own narrative voice.
Requiring that output be vetted and human-jured before before issuing a verdict, the project, meant to ease the administrative case-load, is expected to erode public trust in the courts and run the risk of predisposing judgment before thorough research, influencing the disposition. The makers of the bespoke large language model touts that it is already being used by a few other jurisdictions and has extensive guardrails to prevent hallucinations and inventing precedent, related cases cited with hyperlinks in a fact-checking protocol referred to as Deep Verify. There is presently no requirement for disclosure for rulings adjudicated at with the help of AI.