Colombian judge uses ChatGPT in drafting court ruling

The judge's use of generative AI technology has garnered attention around the world from lawyers and others, and has prompted much debate. The judge, Juan Manuel Padilla, says the tool is a way to more quickly reduce the backlog of cases in Colombia.

Is allowance instantly strangers applauded

A Colombian judge recently turned to the generative artificial intelligence tool ChatGPT to help him issue a ruling in a case, sparking intense debate about whether such technology should be utilized in the courts. 

Juan Manuel Padilla, a judge who presides over the first circuit court in the Caribbean city of Cartagena, told Colombian media outlets that he used the tool to speed up and solidify his decision in a case in which he had to determine whether an autistic child should receive free health services.

“The decision is mine—ChatGPT backs up my decision,” Padilla told Colombian radio program Mañanas Blu

Law.com International has reached out for comment to Padilla.

The judge’s use of generative AI technology has garnered attention around the world from lawyers, tech enthusiasts and others interested in ethics, prompting an animated debate on social media platforms.

The ruling is believed to be the first AI-assisted judgment in Colombia. 

Maria Lorena Flórez Rojas, a Colombian lawyer and assistant law professor at the University of Groningen in the Netherlands, posted a thought piece about the case on LinkedIn asserting that the use of artificial intelligence or any technology within the administration of justice is now a fact.

“It is important to recognize the positive impact that AI systems already have and will continue to have, both commercially and socially,” wrote Flórez Rojas. “However, it is equally relevant that these AI systems can ensure that the risks and other adverse impacts associated with these technologies are managed appropriately and proportionally.”

Technologists warn that biases can creep into AI algorithms and thereby distort court decisions.  

Padilla stated that his judgment, which resulted from a dispute between a health insurance company and the guardian of an autistic child over whether the child is covered for medical treatment, was based on a combination of artificial intelligence and his own determinations.

In his ruling, he detailed the exact questions he posed to ChatGPT, and included the full ChatGPT responses he received. Among his questions were, “Is a minor with autism exempt from paying therapy co-pays?” and “Has the jurisprudence of the constitutional court made favorable decisions in similar cases?”

Juan David Gutiérrez Rodríguez, a Colombian lawyer who teaches public policy courses at the Universidad del Rosario in Bogotá, argues that the ruling in Cartagena “is an opportunity to open a necessary national dialogue about what are the adequate uses of AI in issues where human rights are at play.”

Andrés Fernández de Castro Muñoz, head of the technology, communications and data protection group at Colombian elite firm Gómez-Pinzón Abogados, viewed the judge’s use of AI as “inappropriate and light.”

“Many cases of information bias have been documented in ChatGPT,” wrote Fernández de Castro Muñoz, adding that he would have liked to see additional sources cited in the decision that corroborate the information supplied by the AI tool.

Technology Tools

Padilla explained to Colombian media outlets that legislation passed in 2022, known as Ley 2213, expressly establishes that information technology could be used to speed up judicial decisions in the country.

“Today it might be ChatGPT, but in the future other tools could appear that help judges,” he said.

Padilla said he used the tool to search for precedents and also to draft his ruling more quickly.

The Colombian judge decried the laborious copy-and-paste culture of legal decisions, while simultaneously acknowledging that precedents must be cited in judicial rulings.

Ultimately, he said, it’s the duty of a judge to corroborate information, whether it’s supplied by a legal secretary or via a quick search with artificial intelligence. 

“The application of artificial intelligence in the drafting of judicial texts in no way compromises the ethics of a judge,” Padilla told the Colombian newspaper El Universal.

“At the end of the day, the judge is the one responsible for the decision.”