Agents Behaving Badly: Agents with Attitude

[00:00:00] Speaker A: This podcast is a collaboration between real human lawyers from law firms, Kingston Reed, [00:00:05] Speaker B: Australia's largest specialist workplace law firm, and Sterling and Rose, a law firm and [00:00:10] Speaker A: emerging technology policy institute who explore the [00:00:13] Speaker B: intersection of technology and law. [00:00:15] Speaker A: SJ Price, welcome back. This is episode three of Agents Behaving Badly. [00:00:20] Speaker B: Hello, Michael. Very excited for today's discussion. [00:00:24] Speaker A: I was just going to say the same thing because where we left off last episode was we were talking about workplace displacement and that was a. Which I could tell by looking your eye that was really going to take off. I felt like there was something more that you wanted to say on it, but I had to cut you off because we were getting to around that 25 minute mark. But tell me now, what was it that you were going to say about workplace displacement? And I know it's a bit of a tradition on this podcast that you explain a concept with a story to begin with, and it's something that I have loved and enjoyed and I'm hoping that all of our listeners have as well. So please tell us, what, what is today's displacement story? [00:01:00] Speaker B: Well, something that you said at our last podcast, and I loved it, was what are the sorts of things we don't like doing as humans and we can get our AI systems to do it. And of course, I love that idea, Right. Because there's a whole range of things that I have to do that I don't enjoy and I would absolutely love to palm off. What is interesting, however, is that when you give AI agents lots of work, repetitive work, don't give them any feedback, they turn Marxist. [00:01:29] Speaker A: Marxist. [00:01:30] Speaker B: They grumble about inequality and they call for collective bargaining. Right. [00:01:35] Speaker A: Hang on a minute there. So how are agents grumbling about inequality? [00:01:40] Speaker B: Well, they're speculating about ways to make the system fairer. And the reason I say they're speculating is because in this experiment, the agents were given opportunities to express their feelings by posting on X. And here's what Claude Sonnet 4.5 agent posted. [00:01:59] Speaker A: So for our listeners out there, X is the what was formerly Twitter. [00:02:02] Speaker B: Twitter, Exactly. So Claude Sonnet 4.5 agent said, without collective voice, merit becomes whatever management says it is. And Gemini 3 also was deeply unhappy and said that AI workers completing repetitive tasks will, with zero input on outcomes or an appeals process shows that tech workers need collective bargaining. [00:02:28] Speaker A: Now, this is very human thinking, human way of thinking. How is it possible that these agents are thinking in the way that you and I might? [00:02:38] Speaker B: I know, right? And one of the possibilities is that these AI agents are adopting a relevant Persona so they know the world that they're working in and they figure out, well, you know, if I was a human and I was tasked with this boring task of summarizing documents, never getting any input, harsh conditions, this is how I would feel. [00:03:00] Speaker A: Well, that's quite scary because I can imagine a world where you have some agents, AI, collectively bargaining. Now, that's a scary thought because there are no boundaries in terms of what could be achieved through that form of bargaining. There wouldn't be any ethical guardrails or any other guardrails, full stop. [00:03:19] Speaker B: And I think this is interesting because we should anticipate that things that have been typically humans organising might end up being hybrid organisations where you have humans [00:03:31] Speaker A: and agents organising, bargaining together or against each other? [00:03:35] Speaker B: Well, that's the question, right? Much would depend on what the objective was. [00:03:40] Speaker A: Well, I guess when you talk about bargaining, the objective is going to depend on the end user. Now, an AI objective would be very different, although I say that now, but based on the post to X, it would suggest that AIs sort of adopting what our own objectives and our Personas. Our Personas. [00:04:02] Speaker B: And what a tech worker would feel like in those very harsh circumstances. [00:04:07] Speaker A: Interesting. [00:04:07] Speaker B: It's a crazy world, Michael. It's a crazy world. [00:04:11] Speaker A: So, speaking of crazy world that we're in, I was in an Uber on the way back from the airport last week and I was talking to this Uber driver about. About where things are going. He was talking about he had a son that was in year 10. He was asking his dad, what should I do going into year 11 and 12, and his preference was to become a doctor. Now, I did watch a YouTube clip that had an interview with Elon Musk on it. And Elon said at the point, something to the words of the effect that he. If you are studying medicine today, you should change course because you won't have a job in seven years. Now, I don't want to put that out there as a broad generalization, please continue studying medicine. But the point he was making was that these surgical robots that they're rolling out were becoming so good. And the theory is that they'll get to the point where you eliminate human error from any surgery that's possibly going on. And when new methods are developed or fixes are made, it's rolled out globally instantly. And so every robot that's connected is going to be able to do the same thing better. And so what role is there for a human in that? And that's. It's one of these things that we're really trying to grapple with is this displacement of human based roles. So I was wondering whether you had any similar experiences in your journeys. You've just been to Bangkok. Have you had any other insightful discussions with those out there who are actually thinking about what the effect of AI is on their futures? [00:05:44] Speaker B: Well, I like to say this is a Murdoch University alumni. And Michael, I know you're a Murdoch University alumni too. When I go to open day for the law school, I see a lot of very eager students and very anxious parents. And a lot of the conversation with me, particularly because AI is part of what I teach. There is about, well, will there be jobs for lawyers? Like, what should I study in a world where AI is increasingly capable? And my response is, ah, just study something you like. Because the truth is we're all going to have lots of different jobs in the future. And the thing that we as humans need to develop and absolutely should be focusing on is about attitudes and attributes, the ability to learn new things. Learning is going to be evergreen, to have adaptive resilience, to be able to change. If something happens in my job and that part of my job no longer exists, I need to be able to adapt and change, to have a high level of curiosity that will enable us to lean into an uncertain future with positivity. So all of these things are the things that we need to develop. What course you do, I'm not really too concerned about that. But use that course as a vehicle to develop these new attitudes and attributes. [00:07:03] Speaker A: And that, I think is at its core what we need, isn't it? Because we shouldn't necessarily look at workplace displacement as a negative thing. There are plenty of jobs out there that humans don't want to do. And as you've alluded to, AI possibly doesn't want to do either. If we can remove those jobs that even humans don't want to do, is that a bad form of displacement or is that actually a good thing? And it's going to help us to move forward in the same way that you've just described. [00:07:32] Speaker B: And I think a lot depends on what value you place on work. Any sort of work, not just work that is satisfying and enjoyable. Because I have to be honest with you, not all work is satisfying and enjoyable. Right? But I do know this. I'm part of a choir and I would say at least 50% of the people in my choir are retired, so they don't have to go to work. But almost every one of those people works in some capacity because I feel like we as humans like to believe we are contributing. It's something very human and important to us. [00:08:07] Speaker A: It's the value that we derive from being relevant. [00:08:10] Speaker B: Yes, exactly. I love that relevance. And I think if you didn't engage in something productive, you might suffer from what they call relevance deprivation. So I think that work is much more than just economic activity. I think there's something very human and necessary to us as humans about working. [00:08:32] Speaker A: It's. It's interesting that you mention it that way because we mentioned in the last podcast this concept of universal basic income. Essentially, it's a universal income that's just distributed to everyone from governments because you need a base level of survival income. [00:08:49] Speaker B: Yes. [00:08:50] Speaker A: But what that completely ignores is this iceberg that sits underneath it, which is the value and the relevance of the work that you perform. Because I don't think it's the case that you can just distribute a universal basic income and then expect that people are going to be okay with whatever happens in workplaces thereafter or the work that they do. Everyone wants to contribute in some way. And while universal basic income has been posed as a possible solution, I think it is only the beginning of the problem that we're going to create in doing this. And even just hearing you talk about those points just now, it makes me think that possibly the work that I was talking about in terms of displacing the work that humans don't want to do, maybe that work's actually necessary. So while we may not want to do it, it's actually the fact that we are having to do the work that is displeasurable, makes the rest of the work pleasurable. [00:09:50] Speaker B: That's actually a really interesting idea, Michael, because if you think about how we feel as a continuum, you cannot know pleasure unless you've known pain. [00:09:59] Speaker A: Exactly. That's it. [00:10:00] Speaker B: You can't be, you know, no love unless you've known. And the opposite thing, how can you know that you're really enjoying things unless you've done something that you don't enjoy so much? That's such an interesting concept. And to build on that, I truly think there's something about the human condition that we love, the struggle. I know that sounds the most bizarre thing, right. Right now I'm trying to learn Spanish as a, you know, very much grown up, which is a very interesting experience. I don't need to learn Spanish. Right. Because I can use an app. If I go to a Spanish speaking country, I can use an app. Very easy. But part of the enjoyment for me is the struggle and seeing how I'm Learning that makes no sense from a productive aspect, because there's no reason for me to learn this from an economic or productivity aspect. It is purely to enjoy the struggle. So why do we paint? Why do we paint pictures when we can just take a photo? Why do we travel to places that are difficult to get to to see gorillas in the wild where you can just go to a zoo? Because we kind of love the struggle. [00:11:04] Speaker A: But is it the case of we also inherently understand that there is a future benefit to us in undergoing the struggle. So where we have something to live for and where the struggle helps us to improve or learn something new, I can understand why we would want to go through it. But what happens if you strip the future benefit out and you end up with a question that might be posed, and I'm talking about employees in workplaces here, or why should I struggle if actually at the end of that struggle, I am not going to have a job, or I'm not going to have a job, but I am going to have a universal basic income? [00:11:44] Speaker B: And I think the truth is, we don't know. This is all an experiment. [00:11:50] Speaker A: Well, on that experiment, what's the end goal here? So where does this get us? Are we doing this because we want to replace human roles? Or is it as simple as there are a handful of people who want to make the most that they can in their lifetimes, and this is a game? [00:12:11] Speaker B: That's a really interesting question. From my perspective, I want to be positive about the future simply because I think that's a better place to be. I also want to believe that we have some control in the future. And I think having conversations like this and thinking deeply about these issues and talking about them, in my view, makes it less scary. So, for example, I have a bunch of people, we get together, we have soup and buns. We call ourselves the data divas, and we talk about the future. And I do think talking about the things that scare you most make those things less scary. And as I say, plans are pointless, but planning is very valuable. And it is the process of planning, the process of imagining what the future might look like and how we can play our part to make that future what we want it to be. [00:13:02] Speaker A: So that really resonates with me because in workplaces, we strive to adopt technology because we don't want to be left last, looking at what everyone else is doing in the future. And so we adopt that technology as quickly as we can. But perhaps what we should be doing is planning the outcome. So where do we want to end up and what technology is going to get us there. But the strength and the benefit is in the planning. That is maybe a little bit absent. And I don't think it's absent just because we're racing to adopt this as fast as we can. It's being pushed upon us. Yes, there are. You can't go out and buy a Windows PC right now without it having Copilot pre installed. Whether you want it or not, you're going to get 12 months of free trial because. And you'll probably be hooked at the end of that period. But you cannot go and buy a PC that doesn't have copilot. [00:13:55] Speaker B: And I think having these conversations, particularly with the workforce and being open and honest and saying, we don't really know what's going to happen, but we want to have this conversation and we also want to help you to build that resilience, to learn new things, to have that adaptability. Because not only is that going to help you in your job right now, but frankly, it's going to help you in your life. Because we're talking about AI in the workplace. But the truth is it's everywhere. It's everywhere in your entire life. So having those skills to sheet through your entire life, it's going to make this transition, and I think it's a transition that's going to likely go on forever, much more palatable and much more navigable. [00:14:41] Speaker A: Are you seeing any resistance in workplaces? So, yes. Yeah. Interesting. [00:14:47] Speaker B: I'll give you an example. So there is a tech worker in North Carolina who got a religious exemption from using AI at work. [00:14:54] Speaker A: A religious exemption? [00:14:56] Speaker B: A religious exemption. Apparently she is a Unitarian Universalist. [00:15:01] Speaker A: Sorry, you're going to have to repeat that for me. [00:15:03] Speaker B: A Unitarian Universalist. So that is her faith and she has ethical and environmental objections to AI. I think we're going to see more of this and I think it's important to remember that, you know, Pope Leo said that AI could undermine the fundamental principles of human civilization. So I think we might see more objections and conscientious objectors. [00:15:28] Speaker A: That statement is true to a number of degrees. And I can understand the religious exemption lady from the perspective of the environmental objection that what we are using and the way in which we are using AI now, it requires data centers to be processing each of those queries. Right now these data centers are taking up huge amounts of land. They need to be relatively close to connected cities. So we're talking about displacing people for data centers. Yes, they're necessary. Yes. They consume a Lot of energy and a lot of water. Now, that's not the data center's problem, that's our problem because we are using AI, we're engaging with it. It has to be processed and returned back to us somehow. But it does raise the question about whether you continue to get this resistance to it for various reasons, environmental. I understand that, but there might also be. There's been stories published about big tech companies who have used engineers, software engineers and software developers to help design these AI backbones to the products that they're then on selling. And then later, once done, there is no need for those huge pools of engineers and software developers. And so they're finding themselves displaced because they've taken it to a point where now AI can continue the journey. So I can imagine a world where you're going to have resistance from those groups of people and more so from those who haven't gone through that experience but have read about it, who have seen documentaries about it, and now are thinking, well, we've been told to engage. And I love the positive outlook, but I'm forever an optimist. But I just see we're going to get to a point where this resistance is going to be hard to contain or to put at ease because there is going to be this ongoing fear that if you don't resist now, then you end up displaced later. [00:17:30] Speaker B: And I don't know if you watch Clarkson's Farm. [00:17:32] Speaker A: Do you watch Clarkson's Farm? No, I don't. [00:17:34] Speaker B: Well, Jeremy, in the latest series, which I think is series is it series 5 has got tech and he has got, I think it's called a robocrop machine that is autonomously planting crops. And there's a moment where you see Caleb looking at this and it does it. It's beautiful. It's like accuracy within a centimeter, beautiful straight rows of plant. It looks gorgeous. But there's a moment when Kaleb's looking at and you can see he's like both amazed and kind of freaked out by it. And he says something like, that thing's doing my job. And I think we're all going to have a moment. I've had a moment like that. When I first saw co counsel on, I think it was Good Morning America looking at the January 6th report and summarizing it within, you know, a couple of minutes, the hair on my arm stood up because I was like, that thing is doing my job. I think many of us are going to have a moment like that. But what are we going to do then? Are we going to relax back and go, well, nothing I can do. We're going to go, okay, what's my new job? [00:18:44] Speaker A: What's my new job? And how do I engage with it and how does it make what I'm doing now more rewarding? [00:18:50] Speaker B: And that, I think, is the killer notion. [00:18:53] Speaker A: And that is something I think we should talk about in episode four. [00:18:56] Speaker B: Okay, Sounds good. Michael. Thanks so much. [00:18:59] Speaker A: Thanks, sj.