Andrew Douglas: Our main topic. And for you and I, Kim, in our organization, we use AI for excellence. In other words, we’ll take an idea after using our skills and hone it and test it against AI, and that’s our prescription internally. Not to go and learn something, but to actually hone the skills we have.

Kim McLagan: Yeah.

Andrew Douglas: But daily, we receive complaints, claims, not about us-

Kim McLagan: No, no, no, no.

Andrew Douglas: But claims which are completely crafted in AI. And I thought it was timely we actually dealt with what happens when that occurs, because here is a case in De Bruin where there’s this acrimonious relationship between De Bruin and a caretaker and a committee of management. They didn’t behave well to her, because she really gave them the shit, and you can feel, it doesn’t say that, but you can really tell it, you know? But they also were just worn down by the volume of complaints crafted in AI, as was the commission itself.

Kim McLagan: Yes. And we’re seeing that more and more, aren’t we?

Andrew Douglas: But-

Kim McLagan: Yeah.

Andrew Douglas: The difficulty is, and this is exactly what was done in this decision, you can’t ignore it.

Kim McLagan: Mm.

Andrew Douglas: So that-

Kim McLagan: And that was the problem.

Andrew Douglas: That was the problem.

Kim McLagan: Yeah.

Andrew Douglas: They got to a stage where they go, “I’m not putting up with this rubbish that comes through the door.” But if there’s a kernel in there that’s actually true, you must address it.

Kim McLagan: Yeah, that’s right.

Andrew Douglas: And I think for all of us with AI when we’re dealing, as I said, there’s two ways you can deal with AI. One is do what we do, which is to pursue excellence. But for others, it is to collect together ideas that they don’t have themselves. And you get this bucket of rubbish thrown to you.

Kim McLagan: Yeah.

Andrew Douglas: But it came from a reason, and the reason was legitimate.

Kim McLagan: Yeah.

Andrew Douglas: And so when you’re dealing with people who are using AI, please don’t forget that you must respond, otherwise you end up with workers’ comp claims, stop bullying claims. All these claims, and they will be verified because you failed to address in a reasonable fashion a legitimate complaint. In all that flora and fauna that’s dashing about in front of you that’s nonsense, there is a legitimate complaint. And I thought, we’ve struggled with it, haven’t we? You know, every now and again you get a GP claim or I get a-

Kim McLagan: Oh, and you get pages and pages and you just got to sift through it and it takes forever.

Andrew Douglas: Yeah.

Kim McLagan: And at the end of the day, there’s very little to it.

Andrew Douglas: Yeah, and sometimes nothing at all, and that’s the difficulty. But this committee of management just got so frustrated, they just said “No more.”

Kim McLagan: They just refused to engage. But as you say, there were legitimate complaints that they didn’t address, and so therefore her stop bullying applications-

Andrew Douglas: So the reason I’ve stuck this on is, it is common as humanity that people fall out with each other. It’s very common that an employer and an employee will have disagreements, and the employee has no tool. They don’t have an HR manager, they don’t have in-house legal. Their only tool is AI.

Kim McLagan: Yeah.

Andrew Douglas: So when you’re being resentful and angry, and I am resentful and angry when I read the crap, but you realize that’s the only place they got to go, and so you’ve got to take it seriously. And I know it’s pages of stuff, and it’s headings, and it’s graphs, and it’s a whole lot of stuff, which is irrelevant. But if you don’t do it, you’re going to be into all the thickness of workplace law. And they’ll be criticized for using AI, but it is their only tool. You got to respond. So that’s really the heart of that. Why don’t we get in the case study? We go quickly, don’t we?

Kim McLagan: It’s good.

Andrew Douglas: Yeah.

Kim McLagan: “They had not got along for a long time. Julia rolled her eyes at Denise’s comments about her team’s declining profits, which Denise blamed on Donald Trump’s war and rising fuel costs. At the end of the discussion, Denise looked at Julia, smiled and said, ‘The oil companies may have paid Trump a fortune, but boy, has he cost them. Everyone, even you, Julia, are buying hybrids and EVs.’”

Andrew Douglas: I thought it was for you.

Kim McLagan: I did think of you while I did that. I’ve just got a new Tesla. I love EV, my car. She’s amazing. But anyway. “That was as much as Julia could bear. She huffed, stood up, yawned, and said somewhat bitterly, ‘When you have something better than click bait to explain your drop in performance, come back to me.’”

Andrew Douglas: Good line?

Kim McLagan: Yeah, good line. “Julie was sick of it. She went to her laptop, opened her private Claude account, and entered several prompts about inflation, oil prices, the war, and the apparent disconnect between these factors and the decline in professional service fees. She then asked Claude to tear apart Denise’s argument after entering all of her financial performance figures. The information Julia entered included personal details, client names, and earnings. Seconds later, Claude generated a compelling letter containing tables and references that argued Denise’s position was deeply flawed.

Julia added some further more provocative commentary and then emailed Denise, criticizing her argument and demanding another meeting the following day.” I think Julia’s a bit petty. Anyway, “Denise read the email and was shocked. She replied that she did not feel safe meeting with Julia alone and copied HR into the email. At 8:00 AM the following day, Julia sent a message scheduling a meeting for 1:00 PM. She invited Denise to attend with a support person and copied HR into the message. As the day progressed, Denise became increasingly distressed. At 11:30 AM, she messaged HR to say that she was leaving to see a doctor. Approximately three hours later, HR received a Certificate of Capacity.”

Andrew Douglas: Not an unusual story, really.

Kim McLagan: No.

Andrew Douglas: Okay, the question, did Julia breach privacy law? And the answer is absolutely.

Kim McLagan: Yeah.

Andrew Douglas: The entry of data into an uncontrolled AI process means it is available to other people. So personal names, financial details, addresses, dates of birth. It’s a fundamental breach of privacy, and we’re going to see some very significant arguments around AI in this process.

Kim McLagan: Question though, can she be personally liable for the breach or only the employer?

Andrew Douglas: Employer liable. She can be personally liable under the tort of privacy as well.

Kim McLagan: Right.

Andrew Douglas: So somebody-

Kim McLagan: So not under the Privacy Act then?

Andrew Douglas: Oh, good question. I don’t know completely the answer, which is why I tried to baulk it. I’ll come back to you.

Kim McLagan: Okay. So would Denise have a compensable claim? Yeah, she would.

Andrew Douglas: Yeah, yeah. She’d get up, in any jurisdiction she’d get up.

Kim McLagan: Yeah.

Andrew Douglas: Did Julia bully Denise?

Kim McLagan: Well

Andrew Douglas: Was it repetitive?

Kim McLagan: Well, it depends on when you say they hadn’t gotten along for a long time. It really depends what that conduct was. But that single incident I don’t think would constitute bullying.

Andrew Douglas: But you’ve got three. You got a couple of comments, you’ve got an email, then you’ve got a-

Kim McLagan: Oh, okay. Yeah, yeah, right.

Andrew Douglas: Yeah, I agree. So I guess I’ve asked the question because I’m not sure that it gets that far, but it’s clearly misconduct.

Kim McLagan: Yeah.

Andrew Douglas: And that’s what everyone forgets about bullying is you may not get to the threshold of bullying, but your misconduct is actionable all the way through. Okay. Did Julia in the business breach their safety obligations towards Denise? And if so, what was the breach? They fundamentally breached. They failed to provide a safe place of work. They failed to monitor her health. It’s not bad enough to get to the criminal, the criminal parts, but clear primary breach of duties.

Kim McLagan: And Julia also.

Andrew Douglas: Yeah, yeah, yeah.

Kim McLagan: Section 25 breach.

Andrew Douglas: Yep.

Kim McLagan: And finally-

Andrew Douglas: Look at you in safety. Ooh!

Kim McLagan: I’m learning.

Andrew Douglas: Fail to exercise reasonable care, bring harm to another, and follow process. Can we just quiet, just a quiet moment for Kim. Very good. Okay, number five, did Julia commit acts of misconduct? If so, what should be the consequence for what she has done? She did.

Kim McLagan: Did, yeah.

Andrew Douglas: What do you reckon? What’s the consequence?

Kim McLagan: Well, they should investigate obviously to see just how bad the breach was. But if it’s a blatant breach of an IT or AI policy, then probably termination.

Andrew Douglas: Yeah.

Kim McLagan: Otherwise, first and final warning.

Andrew Douglas: Yeah, I don’t think you get less than first and final, and also her conduct was harmful. It’s clear psychological hazards.

Kim McLagan: Yeah.

Andrew Douglas: It’s damaged. I reckon Julia would be lucky to keep her job, just a general law and safety law.

Kim McLagan: Yeah.

Andrew Douglas: But yes, if she’s breached policy and procedure, and I guess for us, that’s what we’re trying to say to you. If you do not have policy around the use of AI in your business, you are making a very big mistake. You must.

Kim McLagan: Yes.

Andrew Douglas: Well, Kim, I loved catching up with you.

Kim McLagan: Yeah, that was-

Andrew Douglas: I love that. Good game’s a quick game. We’ve nailed it very quickly, but it was fun to catch up and fun cases today as well.

Kim McLagan: Thanks, Andrew.

Andrew Douglas: Okay. See you later guys. Bye-bye.

Kim McLagan: See you. Bye.