There is an old saying that assumption is the mother of all mistakes. In legal practice, we see evidence of that almost daily.
Someone assumes an AI-generated answer is correct. Someone assumes a verbal agreement will be honoured. Someone assumes a property transfer will simply “work itself out”. Someone assumes litigation is the only way to resolve a dispute.
Then reality arrives.
The law, at its heart, is really about certainty. It exists to create clarity where uncertainty lives. It records agreements, protects rights, allocates responsibilities and provides a framework for resolving disagreements when people remember things differently – which, as it turns out, happens rather often.
In this edition of the Commercial Law Report, we explore a number of areas where certainty matters more than ever.
We begin with artificial intelligence. AI has rapidly moved from novelty to necessity in many businesses. It can draft, summarise, analyse and produce impressive-looking answers in seconds. But as recent events have shown, speed and confidence are not the same as accuracy. AI is a remarkable tool, but it remains exactly that – a tool. The responsibility for judgment still rests with the person pressing “send”.
We then turn to a topic that may sound technical but has far-reaching consequences: the requirement that certain agreements must be in writing. A recent legal challenge has raised questions around one of the cornerstones of South African property law. While the debate will ultimately play out in the Constitutional Court, it serves as a useful reminder that written agreements are not administrative red tape. They exist because memories fade, circumstances change and disputes have a habit of appearing when money is involved.
Finally, we look at the recent launch of the Western Cape Society of Mediators which reflects the growing recognition that not every disagreement belongs in a courtroom. Businesses are increasingly seeking faster, more practical and more cost-effective ways to resolve disputes. In many instances, the smartest outcome is not winning the fight – it is finding a way to avoid it.
The common thread running through all these topics is simple. Good decisions are usually made when there is clarity. Bad disputes often begin when there isn’t.
Or, put differently: before you make any assumptions, make sure you know exactly what you’re agreeing to and write it down!
Oh, and before you go – if you’d like to stay in the loop, follow our LinkedIn page for regular legal insight, opinion and updates.
Yours in law
PJ Veldhuizen
TREND TO WATCH:
When AI Invents the Law
Artificial intelligence may be transforming the legal profession, but recent events have highlighted a growing problem: AI’s tendency to confidently present fiction as fact.
In one of the most widely publicised examples, two New York attorneys were sanctioned after submitting court papers containing AI-generated case law and legal citations that simply did not exist. When questioned by the court, the lawyers initially relied on the information provided by the AI platform, only for it to emerge that the authorities had been entirely fabricated.
While the incident made headlines because of its novelty, it exposed a much broader issue facing professionals across industries. AI tools are designed to generate persuasive answers, not necessarily accurate ones. They can summarise, draft and analyse information at remarkable speed, but they are also capable of “hallucinating” facts, sources and legal authorities that appear entirely legitimate at first glance.
More recently, similar incidents have surfaced internationally, prompting courts, regulators and professional bodies to issue warnings around the use of AI in legal research, drafting and decision-making. The message has been consistent: AI may assist professionals, but it cannot replace professional judgment.
The concern extends beyond legal research. Businesses are increasingly using AI to draft contracts, review policies, prepare reports and even formulate strategic decisions. When those outputs are accepted without scrutiny, errors can quickly become liabilities.
The lesson is simple: the faster the technology becomes, the more important human oversight becomes.
My take:
“The problem isn’t that AI makes mistakes. Humans do too. The problem is that AI makes mistakes with absolute confidence. The obligation to verify remains exactly where it has always been – with the professional signing the document. AI is an extraordinary tool, but it has no professional duty, no ethical obligation and no accountability. The moment we forget that, we stop using technology as an advantage and start treating it as a substitute for judgment. That’s where the real risk lies.”