Artificial Intelligence (AI) and the Law in South Africa

By Elize-Mari Gerber · Published

From our article archive. The original wording and publication date are retained; please contact the firm for advice on your current circumstances.

Artificial Intelligence (AI) is no longer just for big tech companies. Whether it’s the chatbot on a customer service website, facial recognition in security systems, or software that screens employees, AI is becoming part of every person’s daily business and professional life, which necessarily include the legal implications of using AI.
But as AI becomes more common, the legal questions around it are getting louder. Can a machine make legally binding decisions? What if it makes a mistake? And who is responsible when it does?
South Africa does not have dedicated AI law yet, but that doesn’t mean AI operates in a legal vacuum. Our current laws already touch on many of the risks and responsibilities that come with using AI.
AI systems rely on data - often personal data. Under the Protection of Personal Information Act (POPIA), any use of that data must be lawful, transparent, and secure. If you’re using AI tools that process customer or employee data, you must:
• Tell people what data you are collecting and why;
• Acquire consent when required;
• Protect that data from misuse;
• Avoid fully automated decisions that significantly affect someone - unless there is a way for a human to review them.
So, who is liable when AI gets it wrong? That is arguably the biggest grey area when relying on AI. Our legal system is based on holding people accountable, but AI makes decisions without direct human input, at least after a certain point. This is the very nature of AI. If an AI system causes harm or loss, we could be looking at a complex web of liability involving developers, users, and data sources.
Even without specific AI laws, South African businesses and professionals should already be ensure that the following fail safes are used:
• Review AI systems for compliance with POPIA;
• Keep records of how automated decisions are made;
• Offer people the right to question or appeal those decisions;
• Stay informed as regulations evolve.
AI can help businesses work smarter - but it also comes with legal risks. The rules around it are still developing, but the responsibility to use it fairly and lawfully already exists. The failure to do so, may give rise to a claim based on negligence if it meets the required standards of our current law and causes direct damage to a person.
Article by Elize-Mari Gerber, Candidate Attorney at Malan Lourens Viljoen Inc.
This article is a general information sheet and should not be used or relied on as legal or other professional advice. No liability can be accepted for any errors or omissions nor for any loss or damage arising from reliance upon any information herein. Always contact your legal adviser for specific and detailed advice.