AI Hallucinations Abound

Since the advent of generative artificial intelligence and its ever-evolving use in the legal research arena, there have been troubling episodes of hallucinations with fake case citations catching both lawyers and judges off guard. We started scratching our heads a couple of years ago following the wide public release of large language models and a well-publicized federal aviation case resulting in $5,000 of sanctions. Dozens more expensive examples of false statements to courts followed, though recently America's largest injury firm was only hit with a grand. Some got pleadings stricken, public reprimands, disqualification or joint and several sanctions under Rule 11, while others were referred their state bar for disciplinary proceedings and of course, all were embarrassed. This summer, a state appellate court even caught a trial judge who decided a case based on fake law and a federal judge in a securities case had to be told by a white shoe law firm that her opinion contained hallucinated cases resulting in a withdrawn opinion. As Supreme Court Chief Justice John Roberts wisely said, "Any use of AI requires caution and humility." It is interesting to note that we are not alone in the world with this problem, though the USA leads with at least 136 such incidents to date. Some counsel fared better by being contrite and falling on their proverbial sword, being bench-slapped with smaller sanctions. Many more seemed rather recalcitrant and argued mistaken briefs were filed or that their work product (despite being vendor checked through subscription with well-known legal service providers) should somehow save them from the same fate. Now these cases have been conveniently catalogued-- on an almost daily basis-- by Damien Charlotin at this handy hallucination link-- https://www.damiencharlotin.com/hallucinations/