South Africa’s Legal Practice Council (LPC) is developing an AI policy to regulate how lawyers use artificial intelligence, following two recent court cases where fabricated case law—generated by AI tools—was submitted as precedent.

The cases, Mavundlela v. KZN MEC for Cooperative Governance and Northbound Processing v. SA Diamond Regulator, were referred to the LPC after legal practitioners cited non-existent judgments in their filings. The LPC, which regulates attorneys, advocates, and candidate legal professionals, considers this a serious breach of conduct.

“This could undermine the integrity of the entire justice system,” said LPC deputy chair Llewellen Curlewis. “Every AI-generated output must be verified before being included in legal submissions.”

Curlewis said the council is working with experts and IT specialists to draft enforceable guidelines. Although AI is still new and challenging to regulate, he emphasized that even unintentional misuse may result in disciplinary action, including disbarment.

South Africa’s legal system relies heavily on the doctrine of precedent, making any misuse of AI especially concerning. The LPC plans to consult with academic institutions, which are also grappling with the impacts of generative AI on assessment and research integrity.

“Universities face these challenges daily,” Curlewis said. “We’ll draw on their experience to shape our approach.”

The AI framework aims to balance innovation with ethical legal practice, ensuring that new technologies don’t erode public trust in the rule of law.