Latest / Views and News with Clarence Ford / Artificial Intelligence and Labour Law: decision maker retains responsibility
Transcript
- 0:00Speaker 1: Grant Herman, welcome. It's great to have you with us.
- 0:03Speaker 2: Good morning, Clarence. Great to be here. Good morning to you,
- 0:06Speaker 2: your listeners, and a big happy birthday to your dad.
- 0:09Speaker 1: Ah, thank you. You've been listening for a while then, Grant.
- 0:11Speaker 2: I was driving to work, yeah.
- 0:13Speaker 1: Okay, I don't know if there's anything that you want
- 0:14Speaker 1: to take issue with me on, but let's talk about
- 0:20Speaker 1: the essence of this conversation. It's about The South African
- 0:23Speaker 1: Post Office, SOC Limited versus Fakudi, establishing that decision makers
- 0:29Speaker 1: retain ultimate goal or legal responsibility and cannot delegate rulings
- 0:33Speaker 1: or findings to AI platforms. Let us understand decision makers
- 0:38Speaker 1: retain ultimate legal responsibility. What is at the heart of
- 0:42Speaker 1: that ruling?
- 0:45Speaker 2: Yes. So people bring a problem before a decision maker,
- 0:48Speaker 2: and it's for the decision maker to then resolve the
- 0:51Speaker 2: dispute and give the right answer, ideally. And it is
- 0:55Speaker 2: obviously required of the decision maker to give a rational answer,
- 0:59Speaker 2: an answer that can be challenged. And we are finding
- 1:03Speaker 2: that decision makers are not doing that. There are a
- 1:05Speaker 2: couple of cases that have been cited besides the post
- 1:09Speaker 2: office one. The post office one is very recent. But
- 1:12Speaker 2: going back even to 2023, there have been cases where
- 1:15Speaker 2: decision makers have come up with crazy rulings based on
- 1:20Speaker 2: AI hallucination, which has been generated. And this obviously leads
- 1:25Speaker 2: to professional negligence claims, much embarrassment to them, and probably
- 1:30Speaker 2: very negative impact on their careers. And this does not
- 1:33Speaker 2: only lie at the feet of decision makers, but we
- 1:37Speaker 2: also need to point out that litigants must also be
- 1:39Speaker 2: very careful of how they prepare for a case. And
- 1:43Speaker 2: there are cases where attorneys have prepared a case, argued
- 1:47Speaker 2: a case, which is based on cases that don't exist.
- 1:50Speaker 2: They've quoted case law that either doesn't exist or it
- 1:55Speaker 2: does not support the case that they are arguing.
- 1:59Speaker 1: Yeah, they're not arguing the merits.
- 2:01Speaker 2: Well, they are, but based on false facts. I'm not
- 2:06Speaker 2: an expert at AI, but how it works, it kind
- 2:09Speaker 2: of falls in the gap sometimes.
- 2:11Speaker 1: It falls in the gap. Absolutely.
- 2:13Speaker 2: And it incorrectly falls in the gap at times. And
- 2:15Speaker 2: as a lawyer, you can't afford to do that. You
- 2:17Speaker 2: can't rely on that.
- 2:18Speaker 1: Are you talking hallucinations specifically when AI makes things up?
- 2:23Speaker 2: Yes, that is on the one hand, hallucinations. And then
- 2:26Speaker 2: on the other hand, in labor law, you need to
- 2:28Speaker 2: also be conscious of the concept of fairness. And that
- 2:31Speaker 2: is something which I would never really be able to understand.
- 2:36Speaker 2: We deal with an unfair dismissal claim in the CCMA case.
- 2:40Speaker 2: And then a commissioner, a human being, will sit and
- 2:42Speaker 2: listen to the dispute. And very often, employers and employees
- 2:47Speaker 2: make themselves guilty of doing research using AI to argue
- 2:51Speaker 2: their case. And I've come across this chairing disciplinary hearings.
- 2:54Speaker 2: I've come across it representing employers where the other side
- 2:58Speaker 2: comes forward with arguments that don't match the law at
- 3:02Speaker 2: the CCMA, don't match South African labor law. And that
- 3:05Speaker 2: is incredibly risky when you're coming to a battle and
- 3:08Speaker 2: you've brought the wrong ammunition with you.
- 3:11Speaker 1: Yeah, I would imagine caseload workload for a commissioner is
- 3:14Speaker 1: significant in that maybe AI can assist with that caseload.
- 3:20Speaker 1: You're not arguing against the use of AI. You're saying
- 3:23Speaker 1: be circumspect of how you use it.
- 3:27Speaker 2: Correct. I mean, we use AI without knowing it. When
- 3:29Speaker 2: you're writing email, half the words are underlined in red
- 3:33Speaker 2: or blue. And that's AI telling us that we need
- 3:36Speaker 2: to check our spelling and our grammar. So AI is
- 3:39Speaker 2: very prevalent. If we use the South African Post Office
- 3:43Speaker 2: case as an example, there the commissioner made massive errors.
- 3:48Speaker 2: He quoted and constructed a Constitutional Court case which doesn't exist.
- 3:53Speaker 2: And the funny thing is one of the litigants in
- 3:55Speaker 2: that Constitutional Court case was the South African Post Office
- 3:58Speaker 2: in this hallucinated case. So this is a case where
- 4:03Speaker 2: the post office had brought a jurisdictional ruling before the
- 4:07Speaker 2: labor court, which the commissioner got wrong. And the commissioner
- 4:10Speaker 2: relied upon this hallucinated South African post office case in
- 4:14Speaker 2: his decision. And that case went completely contrary to the
- 4:19Speaker 2: law as we understand it. And a very basic understanding
- 4:23Speaker 2: of the law would have told him, hey, this is
- 4:24Speaker 2: very strange. I need to research this a bit more deeply.
- 4:28Speaker 2: I need to go and look for that case. I
- 4:29Speaker 2: need to read it.
- 4:30Speaker 1: Absolutely.
- 4:30Speaker 2: I need to see what the circumstances were there in
- 4:32Speaker 2: that case. Cross-reference. And that case could not be found. Absolutely.
- 4:36Speaker 1: Yeah, so that is an omission, an oversight on his
- 4:39Speaker 1: part and correctly sanctioned. But we've also seen many businesses
- 4:45Speaker 1: use AI for automated maybe recruitment screenings. There's performance tracking.
- 4:50Speaker 1: as maybe selecting candidates for retrenchment as well. I'd like
- 4:54Speaker 1: to know, just on the other side, how do employers
- 4:56Speaker 1: ensure those algorithms don't inadvertently, you know, trigger the unfair
- 5:01Speaker 1: labor practice or discrimination, you know, kind of claims under
- 5:05Speaker 1: the LRA especially?
- 5:06Speaker 2: Yeah, it's important to speak to a human being about it.
- 5:10Speaker 1: Yeah.
- 5:10Speaker 2: Unfortunately, we're still there. We're still at the place where
- 5:13Speaker 2: you need to bounce off a human being. Employers approach us,
- 5:17Speaker 2: our members approach us often with a scenario And then
- 5:19Speaker 2: they will tell us, this is what I found, this
- 5:21Speaker 2: is what I researched. And very often it's wrong. And unfortunately,
- 5:26Speaker 2: it seems as though AI gives you the answer you want.
- 5:29Speaker 2: And that is very dangerous for any litigant. And it
- 5:33Speaker 2: goes for the employee as well. If you put punch
- 5:35Speaker 2: in the facts and you word it in a certain way,
- 5:38Speaker 2: AI is going to tell you you've got a great case.
- 5:40Speaker 2: And you're going to come forward with that. And unfortunately,
- 5:43Speaker 2: you're going to be caught out. So it's risky to
- 5:45Speaker 2: employers and to employees. Absolutely speak to a human being.
- 5:48Speaker 2: Don't rely on the AI.
- 5:50Speaker 1: Yeah, I don't know if you see that boundary between
- 5:54Speaker 1: AI as a support tool versus a decision maker. And
- 5:59Speaker 1: I'm wondering if that boundary is clear.
- 6:03Speaker 2: Yes, I mean, you're welcome to use AI. And I
- 6:06Speaker 2: probably encourage people to use AI to research, but you
- 6:10Speaker 2: need to verify the authenticity of that information manually by
- 6:15Speaker 2: going through it yourself.
- 6:18Speaker 1: You know, if the court is also not necessarily being
- 6:22Speaker 1: meticulous in its evaluation of the merits of that particular case,
- 6:27Speaker 1: they can get a drawing. And how do you spot AI?
- 6:29Speaker 1: How do you spot a hallucination? Somebody's double checking then. Yes.
- 6:36Speaker 2: I mean, when you're litigating, the other side is definitely
- 6:38Speaker 2: going to be checking your references. So that is something
- 6:43Speaker 2: to be careful of. If you're quoting a case in
- 6:46Speaker 2: your argument, they're going to be looking at that case
- 6:48Speaker 2: and dissecting it and looking to see why it doesn't
- 6:52Speaker 2: apply in this case. So that's when AI will come out.
- 6:56Speaker 2: But it's for the other side then to argue that
- 6:59Speaker 2: and the judge to come to the correct conclusion.
- 7:03Speaker 1: Well, I think we're flagging again how we need to
- 7:07Speaker 1: treat AI. And I think you've given us some really
- 7:10Speaker 1: important insights there. Grant Herman, spokesperson from NISA, we appreciate
- 7:16Speaker 1: your time. And I think that sweet spot resides right
- 7:21Speaker 1: there between AI as a decision support tool and an
- 7:28Speaker 1: autonomous decision maker, especially when it comes to the law.