South African Judge Investigated Over AI-Generated "Hallucinations" in Court Ruling

An acting judge in the Johannesburg High Court, Themba Khaba, is under scrutiny by the Legal Practice Council (LPC) for allegedly using artificial intelligence to draft a judgment. The ruling in question, which concerned a dispute between parents over their children's schooling, contained at least 11 "hallucinations" – citations of non-existent or irrelevant case law. While the appeal court upheld Khaba's ultimate findings, Judge Ingrid Opperman issued a separate ruling highlighting these discrepancies and suggesting AI as the most plausible explanation.
Judge Opperman's ruling meticulously detailed the 11 instances of incorrect citations, which were brought to the court's attention by the appellant's lawyers and confirmed by a senior librarian. She emphasized that while she made no definitive finding on AI use, the implications for independent judicial reasoning were significant. Opperman noted the lack of precedent in South Africa or most other countries where a judge, as opposed to lawyers, has been found to use AI in rulings, contrasting it with Lord Justice Colin Birss of the England and Wales Court of Appeal who publicly disclosed using ChatGPT as a drafting aid, subject to verification.
The case has prompted discussions within the South African judiciary about developing an AI policy to leverage its benefits while mitigating risks to justice administration. Opperman underscored that non-existent cases and jumbled quotes do not constitute law, and suggested that if the reasoning were truly AI-generated, it would infringe upon the litigants' rights to be heard by a human court. However, she stressed that Khaba must be afforded due process.
Crucially, the ruling also highlights a gap in accountability for acting judges in South Africa. The Judicial Service Commission (JSC) lacks direct authority over them. Opperman concluded that the LPC, which has disciplinary jurisdiction over advocates and attorneys, is the most effective forum for investigation. She will be forwarding her judgment to the LPC to initiate an inquiry into Khaba's conduct, which could amount to professional misconduct.
This incident marks a significant moment for the intersection of AI and law in Africa, raising critical questions about judicial integrity, accountability, and the responsible integration of advanced technologies within legal systems. It underscores the urgent need for clear guidelines and policies to govern AI use in judicial processes across the continent.
Source
More in policy
Mauritius Launches Business AI Adoption Survey to Gauge National Strategy Effectiveness
Mauritius has launched a national survey to measure the actual adoption and governance of AI among its local businesses, providing crucial data to assess the effectiveness of its…
AI Fuels Over Half of African Cybercrime Amidst Soaring Financial Losses, INTERPOL Reports
An INTERPOL report highlights that AI is now responsible for 55% of cybercrimes in Africa, leading to a significant increase in financial losses to $484 million. The continent…
Kenya Unveils Ambitious AI Policy with Extraterritorial Reach and Sovereignty Focus
Kenya has published its Draft Artificial Intelligence and Emerging Technologies Policy, establishing a comprehensive governance framework with extraterritorial scope and a focus…
AI Fuels Over Half of Cybercrime in Africa, INTERPOL Report Reveals
A recent INTERPOL report indicates that artificial intelligence is now involved in 55% of cybercrime incidents across Africa, significantly escalating fraud, ransomware, and…
The dispatch
One email a day. The AI stories shaping Africa.
Rewritten for clarity, sourced always. No spam; unsubscribe anytime.



