Circuit Judge Opines on Benefits of AI in Judicial Decision-Making
Summary
On May 29, 2024, Eleventh Circuit Judge Kevin Newsom, in a self-described “unusual” concurrence, suggested that using AI as part of the interpretative analysis could be beneficial. He proposed that generative AI, like OpenAI’s ChatGPT, could aid courts in understanding the “ordinary meaning” of words and phrases in legal documents—such as, in the case at issue, whether an in-ground trampoline would be considered “landscaping” under an insurance policy.
After dictionaries “left a little something to be desired,” Judge Newsom “wonder[ed] what ChatGPT thinks about all this?” and asked a clerk to find out. The result aligned with Judge Newsom’s impression of how landscaping is used “American English in the real world.” Taking the experiment a step further, he tested a few models’ responses to whether an inground trampoline is landscaping—and found those to answer consistently in-line with common usage. As a result, Judge Newsom’s concurrence discussed why large language models (LLMs) may be particularly useful for understanding language (e.g., they are large statistical models of how words are used) and acknowledged the potential risks—but emphasized that AI is here to stay and should be used responsibly alongside traditional methods.
