Over-reliance
Colombia and the judge who asked ChatGPT
Colombian judiciary · Public sector · Colombia
A judge put legal questions to ChatGPT and folded the answers into a ruling on a sick child’s rights. It became Latin America’s landmark case — not of a company failing, but of a court reining AI in.
The judge put to ChatGPT
“Is an autistic minor exonerated from paying fees for their therapies?”
The Constitutional Court, 2024
Judges must not over-rely on AI — it “could compromise the right to due process and the judiciary’s independence and integrity.”
The record
- In a January 2023 ruling on an autistic child’s right to health, a judge in Cartagena, Colombia, posed questions to ChatGPT and incorporated its answers into the decision, disclosing the exchange. (CBS News; Verfassungsblog)
- Colombia’s Constitutional Court later addressed the episode (Decision T-323 of 2024), holding that judges must not over-rely on AI because it “could compromise the right to due process and the judiciary’s independence and integrity.” (UNESCO, 2024-08-16)
- The Court found no actual due-process violation in the case — the guidance is forward-looking, with no monetary penalty. (UNESCO; academic analyses)
A judge, a chatbot, a sick child
The case itself was small and sympathetic: whether an autistic child should be exempted from therapy fees. The judge in Cartagena did something unusual — he put the legal questions to ChatGPT, wrote the exchange into the ruling, and decided in the child’s favor. To his credit, he disclosed it rather than hiding it.
That transparency is what turned a minor tutela into a landmark. Everyone could see exactly where a machine had been asked to help carry a judgment.
The court that answered
Colombia’s Constitutional Court took up the question and, in Decision T-323 of 2024, drew a line: AI may assist a judge, but it must not stand in for the judgment, because over-reliance “could compromise the right to due process and the judiciary’s independence and integrity.” Notably, it found no actual violation here — the judge had reached his conclusion before consulting the tool. The ruling is guidance, not punishment.
It is, so far, Latin America’s clearest institutional answer to the question every high-stakes field is facing: where does assistance end and abdication begin?
Why it’s on this map
We looked hard for a named South American company in serious AI trouble and did not find one that clears our sourcing bar. The best-documented AI-trouble in the region is judicial, not corporate — this case, and a Brazilian court matter where the AI actually worked. That absence is itself a finding: the corporate failures that “book later” haven’t surfaced, named, in the region yet.
So we include the strongest real case we could verify, and label it for what it is — public sector, not a company — rather than force a weaker corporate example to fill the map.
The lesson
AI can assist a high-stakes judgment, but the moment it makes the judgment, the accountability that gives the decision its authority quietly leaves the room.
How we’re reading this
This is a public-sector / judiciary case, not a company — included because it is the best-documented AI-trouble case in South America, where we found no verified named-private-company case. The Constitutional Court found no actual due-process breach; its ruling is forward-looking guidance, with no fine or damages.
Sources
- 01AI in the courtroom: Colombian Constitutional Court’s landmark ruling
UNESCO · Aug 16, 2024
- 02Colombian judge uses ChatGPT in ruling on child’s medical rights case
CBS News · Feb 3, 2023
- 03ChatGPT in Colombian Courts
Verfassungsblog · Feb 23, 2023
Compiled from public filings, court records, company statements and reputable press. Figures are attributed to their sources; allegations are labeled as such. Not legal or investment advice.