Artificial intelligence has officially entered the legal boardroom. It is drafting, summarising, researching, reviewing contracts and, occasionally, making things up with the confidence of a first-year candidate attorney after two espressos.
The issue is no longer whether businesses and law firms should use AI. They already are. The real question is whether they are using it with enough discipline, oversight and legal common sense.
In South Africa, AI regulation remains a work in progress. The draft National AI Policy was published in April 2026, then withdrawn after it emerged that the document itself contained fake AI-generated references. A fairly spectacular own goal, but also a useful reminder: AI may be clever, but it is not accountable. People are.
Globally, the legal sector is moving quickly. AI-native law firms are emerging, promising faster turnaround times, fixed fees and fewer hours spent on routine document work. Legal technology platforms are being embedded into research, contract review, litigation preparation and compliance workflows. In-house legal teams are also under pressure to do more, faster, with leaner teams and better data.
This is not necessarily a bad thing.
Used properly, AI can reduce repetitive work, improve access to information and free lawyers to focus on strategy, judgment and negotiation. But the danger lies in confusing output with advice.
Law is not simply a question-and-answer exercise. It involves context, risk, timing, people, personalities, commercial pressure and, very often, the uncomfortable grey areas where the real disputes live. AI can process information. It cannot read the room.
For businesses, the risks are practical. Confidential information may be uploaded into tools without proper checks. AI-generated clauses may look impressive but fail under legal scrutiny. Automated decisions may create POPIA, employment, discrimination or consumer protection concerns. Cybercriminals are also using AI to impersonate executives, draft convincing correspondence and manipulate decision-makers.
Then there is the question of legal privilege – and this is where things become particularly dangerous. Uploading the details of your dispute, transaction or legal strategy to an AI service provider is not the same as communicating with your attorney and may not attract legal privilege. Put differently, in litigation, that information could potentially become discoverable to the opposition. That is not just inconvenient; it could be a strategic disaster.
The courts have already shown little patience for AI-generated fiction dressed up as legal research. Fabricated case law is not a technological hiccup; it is a professional failure. The answer is not to ban AI. That would be like banning email because someone once hit “reply all”. The answer is governance.
Every business should have a clear AI policy covering what tools may be used, what information may be shared, who must verify outputs, how confidential data is protected, and where human approval is required. Legal teams should also review contracts, employment policies, supplier agreements and data protection processes to ensure AI is not quietly creating liability in the background.
AI is here to stay. It will change how legal work is priced, delivered and managed. It will make some tasks faster and some business models uncomfortable. Good.
But it should be treated as digital staff, not digital counsel. Helpful, fast and occasionally brilliant, but still in need of supervision.
So yes, all in favour say AI. Just don’t let it chair the meeting.