South African Judge Faces Scrutiny Over AI-Generated Judgment

• An acting judge in the Johannesburg High Court faces scrutiny over a judgment that may have contained AI-generated 'hallucinations'.
• The judgment included at least 11 fictitious or inaccurate case citations, which were brought to the attention of the appeal judges by lawyers for the appellant.
• Judge Ingrid Opperman suggested that the most plausible explanation for these discrepancies is the use of AI and 'hallucinations'.
• The incident highlights the need for judges and lawyers to be aware of the potential risks and implications of using AI in judicial decision-making.

An acting judge in the Johannesburg High Court has been ordered to explain himself to the Legal Practice Council (LPC) over a judgment that may have contained AI-generated 'hallucinations'. The judgment, which was recently appealed by three judges, included at least 11 fictitious or inaccurate case citations. In her ruling, Judge Ingrid Opperman detailed all 11 discrepancies, which were brought to her attention by lawyers for the appellant and confirmed by a senior librarian. She noted that in some instances, the correct case name was used but with an incorrect citation, while in others, the correct case number and reference were given but with an incorrect passage quoted. One notable example was the non-existent case of Lubbe v Volkswagen SA. Opperman suggested that the most plausible explanation for these discrepancies is the use of AI and 'hallucinations', a term used to describe AI-generated errors.

This incident highlights the need for judges and lawyers to be aware of the potential risks and implications of using AI in judicial decision-making. The use of AI can lead to 'hallucinations' or AI-generated errors, which can erode independent judicial reasoning. In South Africa, the judiciary is currently drafting an AI policy to take advantage of the opportunities presented by AI while guarding against its risks to the administration of justice. This development also raises questions about the role of judges and acting judges in using AI tools, such as ChatGPT, for research purposes. Sir Colin Birss, the Master of the Rolls, has publicly disclosed that he used ChatGPT to produce a summary of an area of law while drafting a judgment, but emphasized that he only used it as a drafting aid subject to judicial verification.

The use of AI in judicial decision-making raises concerns about the erosion of independent judicial reasoning and the potential for 'hallucinations' or AI-generated errors. Lawyers should be cautious when citing or relying on judgments that may have been generated by AI, and ensure that they properly verify the accuracy of any cited authorities. The incident also highlights the need for judges and lawyers to be aware of the potential risks and implications of using AI in judicial decision-making, and to take steps to prevent its misuse. As Judge Opperman noted, 'the acting judge has to be heard and due process is to be followed in this regard.'

This development highlights the need for judges and lawyers to be aware of the potential risks and implications of using AI in judicial decision-making, including the risk of 'hallucinations' and the erosion of independent judicial reasoning. Lawyers should be cautious when citing or relying on judgments that may have been generated by AI, and ensure that they properly verify the accuracy of any cited authorities.

Source: Original reporting via AllAfrica SA