Chatbot ChatGPT is back in the legal news this week after a Colombian judge has admitted he used the artificial intelligence (AI) tool in a written court ruling.
In an article last week, we highlighted that AI platforms such as ChatGPT are garnering increased attention in the legal industry with AI experts claiming they could soon revolutionise the provision of legal services.
Then, on 3 February 2023, the Guardian newspaper reported that a judge in Colombia noted he used ChatGPT to pose legal questions about a case and incorporated its responses in his decision. Judge Juan Manuel Padilla Garcia, who presides over the First Circuit Court in the city of Cartagena, was ruling on whether an autistic child's insurance should cover all costs of his medical treatment.
Court documents show that the questions Garcia entered into the chatbot included “Is an autistic minor exonerated from paying fees for their therapies?” and “Has the jurisprudence of the constitutional court made favourable decisions in similar cases?” The Judge then incorporated ChatGPT's full responses into his decision.
This is thought to be the first time a judge has quoted an AI tool such as ChatGPT in their judgment, or certainly the first time they've admitted to it. A law in Colombia passed in 2022 emphasised that lawyers should be using technology where they can to work more efficiently. However, Garcia's admission has raised incredulity and concern from commentators around the world over the ethics of using the tool in a court judgment with real world consequences.
ChatGPT has a number of limitations, namely that it hasn't been trained on information beyond the end of 2021, it doesn't explain or give the source(s) of information it provides and it often emits plausible-sounding but incorrect or nonsensical answers. Even ChatGPT's developers have previously noted that the chatbot still has limitations and is not fully reliable.
Given the increasing presence of AI in law, it seems likely stories such as this will only increase in frequency. For the moment, however, it seems clear that using AI tools such as ChatGPT for any significant or consequential decision-making remains inadvisable.
Want to hear more from us? 
