US court uses ChatGPT to deliver ruling

This appears to be the first instance of a legal judgement being made with the aid of an AI text generator as a judge in Colombia, US used ChatGPT to issue a court decision. The case featured a disagreement with a health insurance provider over whether or not a child with autism should be covered for medical care.

According to a court record dated January 30, 2023, Judge Juan Manuel Padilla Garcia, who presides over the First Circuit Court in the city of Cartagena, claimed he used the AI tool to pose legal questions concerning the case and included its responses in his judgement.

These texts created by AI are not intended to take the place of the judge's ruling in any manner. The goal is to reduce the amount of time needed to draft decisions after verifying the data provided by AI, Judge Garcia said.

Also Read: India gets its first ChatGPT-powered AI chatbot, ‘Lexi’

Artificial intelligence (AI) was used to determine the justifications for this choice. So, the court asked a portion of the legal queries raised in such processes, he added.

The use of AI in court rulings is not prohibited by Colombian law, but programmes like ChatGPT are infamous for providing responses that are either biassed, discriminating, or just plain incorrect. This is so because the language model just synthesises words based on probability from the millions of examples used to train the system, not actually understanding the text.

The legal questions entered into ChatGPT included if a minor with autism was exempt from having to pay for their therapies and the constitutional court's precedent resulted in favourable decisions in such cases.

Also Read: CBSE issues orders on use of ChatGPT in class 10, 12 board exams

Judge Garcia apparently broke new ground by admitting to using the chatbot's complete responses in the judgement. The judge added his own observations on relevant legal precedents.

He claimed that ChatGPT had been used to further the case for the chosen course of action. The judge first describes the conversations with the AI before adopting its responses and his own legal justifications as the basis for its judgement.

“ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness,” ChatGPT-parent OpenAI CEO Sam Altman earlier said as people had started questioning how authentic the chatbot’s response was.