"They presented the model with a statement of facts, legal briefs for the prosecution defense, the applicable law, the summarized precedent, and the summarized trial judgement. And they asked the model whether it would support the trial decision, to see how the AI responded and compare that to prior research (Spamann and Klöhn, 2016, 2024), that looked at differences in the way that judges and law students decided that test case."
"Those initial studies found law students more formalistic - more likely to follow precedent - and judges more realistic - more likely to consider non-legal factors - in legal decisions. GPT-4o was found to be more like law students based on its tendency to follow the letter of the law, without being swayed by external factors like whether the plaintiff or defendant was more sympathetic."
GPT-4o was tested on an ICTY war crimes appeal prompt and given a statement of facts, prosecution and defense briefs, applicable law, summarized precedent, and the trial judgment. The model was asked whether to affirm or reverse the lower court and showed a tendency to follow precedent and the letter of the law, resembling law students' formalism. GPT-5 was later used to replicate a study involving 61 US federal judges deciding routine jurisdictional questions. GPT-5 adhered more strictly to legal rules than the judges and was less influenced by non-legal factors. Whether AI is appropriate for judicial decision-making remains unresolved.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]