273 – The Future of Work Beyond AI Empires

[00:00:07] David: Hi, and welcome to the IDEMS Podcast. I’m David Stern and founding director of IDEMS, and it’s my pleasure to be here today with Kate Fleming, a fellow director.

Hi, Kate.

[00:00:16] Kate: Hi, David.

[00:00:18] David: We’re continuing our series, which has come out of the book The Empires of AI, and what is it you wanna dig into today?

[00:00:27] Kate: I think the topic that I am most interested in right now, what is out in the world right now, is a lot of conversation about the future of work. And that future of work model is very tied to assuming we stay the course with AI. So we’ve been talking a lot about how AI might look different, and particularly in an impact context what AI as a useful intervention tool looks like.

And I wanna talk about, well, in this vision, what is the future of work, what is the role of humans in this? I think we see there’s a need for a lot of new skills, new work, different skills, definitely. But I think rather than all the doom and gloom, it feels exciting and interesting. And so I wanted to kind of dig into that.

[00:01:17] David: I guess my perspective on this is that I feel we’re, as a society, we are faced with a choice on what we want the future of work to be. I don’t think the doom and gloom is potentially misplaced, that is a genuine possibility. But it’s not the only way we could choose to have these technologies influence our societies.

My favourite instance to really think about this is education and the education space. So, right now things are changing fast, there’s a lot of disruption in education, there’s a question about can you get better results with AI tutors than you could do in a classroom setting? And I want us to look past that.

Let’s say we are in a future, a hundred years down the line. I think there is no doubt that in terms of the educational outcomes of the past, in a hundred years time AI tutors could help students outperform the education systems of the past. So the question is, in that future, what do we see as the future of education? And the future of education that I see doesn’t look that different from education in some ways in its current form.

I like teachers in a classroom, which means I like the idea that in a hundred years time, the concept of school will still exist, kids will still go to school. Maybe not all of them, maybe there will be exceptions and so on and there’ll be other things that would be possible that are not necessarily as easy now. There are lots of people who are homeschooled now, maybe there’ll be more people who are homeschooled. But I think for the majority of children across the world, actually school is positive and can be positive for more than just the education outcomes.

And in fact, that’s the key. I think teachers and the humans in the classroom will be able to spend much more of their time on human interactions than they currently can because the AI systems will be helping with a lot of the bureaucratic effort that is currently weighing down teachers. That’s my hope for the school, my hope for the school of the future is that it has at least as many humans in it as they currently exist, but that those humans spend a higher proportion of their time on human interactions with students, with other teachers, with the parents. Those are the human interactions which I think it would be great if the educators can invest more of their time in. So that would be looking a hundred years into the future, that’s what I hope we will see.

[00:04:22] Kate: I like that vision and I think it is important because it is getting at that core idea that AI is a tool and if it’s a tool used well, it should free up people to do the things that people do best. And I think of this even when I was in school, a lot of the classroom day is very focused on, maybe one kid is very stuck on something, but the teacher’s very aware they need to make this progress.

It can become this weird social dynamic because you’re trying to simultaneously solve a lot of problems in the context of the classroom that might not be the best use of that social, collaborative, interesting time when things that are quite creative could be emerging. Instead, you’re doing a lot of mundane things, trying to get through things, I just need to see that you’ve mastered this, somebody’s stuck, all of those kinds of things.

So I think what I hear and what you’re articulating is: you can move some of those things, some of those activities that might co-opt classroom time right now and co-opt human labour in maybe ways that are not as interesting, collaborative, exciting, and those things go to AI, those things are handled by AI. And then, yeah, the classroom space becomes quite different.

[00:05:44] David: This might change things even more drastically in the sense that the teacher might no longer be really the source of the knowledge. They might no longer be the presentation of the knowledge, that might be coming through interactions with AI. So what is that role that the teacher is enabling, how is that happening?

These are questions which I don’t have answers to right now. I think there’s a lot of research happening around this, there’s a lot of learning, which we are doing right now. We cannot, or I cannot imagine a hundred years down the line what the exact interactions would be. But what I can do is I can imagine two futures, one, which I feel is desirable, where the nature of education and schooling maybe changes in certain ways, but it’s all about making it more human because of the technology enhancing it.

And then the alternative is really to say that there’s a narrow set of jobs which people are trained for, that education as it currently exists is no longer needed, people are trained by AI systems to do the narrow things that they need to do, and they’re living in what I would consider a dystopian view where, actually, the jobs or the roles that human take are funnelled towards through these AI systems because that’s what you are needed for.

[00:07:08] Kate: And that’s that reverse centaur concept, the computer is the head of the horse and we’re all just horses asses, just kind of being led around by the computer brain, just putting in the grunt work. And so, yeah, that is not the ideal vision. And I did think about it while you were talking, so there’s one future of work, the real doom and gloom, which is that everyone’s gonna be unemployed.

I think we would argue that is, I don’t even think that’s a reality with anything that we’re seeing in existing AI. And actually I did see, oh gosh, I need to remember who says things, one of the big tech leaders, maybe Jeff Bezos, somebody like that, was saying, no, there’s actually going to be so much more work with existing AI.

But to me, a lot of that work is what you’re describing, it’s reverse centaur work where it’s like, now you’re just managing, you’re following behind the AI and cleaning up its messes, you’re constantly trying to hold back the tide of its chaos. And that becomes kind of very managerial work, it’s not interesting, you have no time to be creative because you’re just trying to keep everything from descending into chaos.

And so, if that’s the kind of system you have, then it feels like, oh yeah, that’s gonna require a lot of cleaning and maintenance. I don’t know, it’s very low level kind of work in some way. 

[00:08:24] David: Let’s think about this in terms of actually the systems that get built. A lot of this is about responsibility, who is responsible for what and how is that responsibility playing out in different ways? But I think let’s reign this back in, because I think you said that we’re not gonna get total unemployment. No, but we could get very high rates of unemployment in a lot of contexts where existing jobs that employ a lot of people suddenly one person can do 50 people’s worth of work, and so there can be whole areas, and that could lead to really high unemployment rates in certain sectors.

This is a reality which is possible, and I think one of the things which is interesting there, is that if you think of this, and you think of this in terms of the big tech at the moment, they’ve been firing a lot of people because of AI, despite the fact that they have more and more money.

So now this is a choice of the sort of enterprises we want to actually lead the world in different ways. They’re not firing people because they can’t afford to keep them, they’re not firing people because they couldn’t add value to the company, they’re firing people because they’re trying to optimise, to maximise on certain things.

And this is where that might mean that we need to maximise on different things. Maybe we want companies that are deliberately then trying to make use of human labour in interesting ways and adding value to society through it. Maybe social enterprises would actually take a higher role and take a higher proportion of things because maybe you don’t need that many employees, but is society better off because of having those employees and do they add value? Could we go back to an era where in customer service you could actually talk to a human again?

So it’s a ridiculous thing, but actually a proper human conversation as part of something, I would love that, maybe people would start paying for that as part of customer service. Not because it’s needed, but because it’s nice, it’s something people want.

So maybe we could actually get back to the situation where some of these choices, so if you have really good, much better than currently exist AI systems to resolve disputes and to do all this stuff, then maybe alongside those, you could also choose to speak to a human agent just because you want to. Maybe that’s not the most cost effective, but maybe those companies would be the ones that become rewarded because they’re offering services that people want, not because they have to, but because they want to.

It’s really interesting, so it’s not just the future of work, it’s the future of business. Are we trying to maximise or minimise human input, or are we trying to maximise social value and the company prestige and so on?

[00:11:23] Kate: And I think what you are getting at is the distance between here and then, or what’s implicit underlying, and this is all technological change, technology is anything that’s a tool that changes how things work. But yeah, I mean, I think we’ve talked about this before, but you know, the arc of history is, well, if you were a carriage driver when the car came in, there are always these periods where people are really fighting well, we need to protect carriage drivers, well, actually we don’t really need them anymore.

But then you have people, there are these really hard issues and I think this is where if you’re really focused on policy and people in the immediate and recognising how destabilising things are, it is really hard work to skill people for new jobs. If you’re young, fine, you’ll be fine. But if you’re older and you’re mid-career, it’s not easy to just channel you into new work or to point you in a different direction. You kind of do need to gently move people in new directions.

But I guess what I would say in the spaces that we work in is, this is where value goes to, what gets paid for all of that. A lot of the jobs that people are currently doing, they don’t actually enjoy that much. You are doing it because you have to to get paid. And there are all kinds of unsolved problems that people often work on. They’re volunteering, they do it through church, they do it through some avenue. They really care about this issue and they actually have valuable human skills that get applied in their communities in some issue, whatever the thing is that’s completely underserved. This sounds so idealistic and not the way the world works, but why can’t it work this way? 

[00:13:01] David: I wanna pick out on two things you’ve said. Maybe I’ll start with a second and then I’ll come back to the youth versus experienced issue, because I think AI is flipping that narrative that you had. With technologies in the past, it’s the young people who are fine and the older people who are struggling. Actually, right now, there is a recognition that the youth are struggling and with good reason. And I think we need to call that out as well, so I’ll come back to that in a second.

But I do want to start with your utopian idea that actually maybe more people could be working on things they want to work on in a future. And I think that it is not a totally unrealistic scenario a hundred years down the line. Getting there, there are societal choices that need to be made. We need to understand what it is that we value and how it’s valued to be able to put in place structures to be able to do this.

I come back to very simple things on this that I think are possible to imagine. There’s a number of interesting opportunities where if you think about the interactions that people rarely want in their job, often they’re very human interactions. Not always, but a lot of people enjoy those human interactions, and there is good reason to expect that that is going to be the thing which you cannot replace. And so that could be what humans end up doing more of.

Almost all the bureaucratic work, nobody enjoys doing their expenses, the simple things that need to be done. And these are simple pipelines, which could be made much easier, almost trivial, and AI systems could really help with that. But that human interaction, that really personal piece, it’s not only that I don’t think we want that to be replaced, it’s that trying to replace it would probably actually not be as effective as simply giving people time and space to do it. And that’s what’s so exciting. With what the AI systems are getting really good at, it could free people up to have those human interactions, which would then make the world and the systems work.

So this is not just utopian thinking, but it does need to be designed for, and that’s the thing which I think is so important, that while a lot of people are fighting little minute battles on different things – which I’m not saying they’re not important, they are important – somebody needs to be thinking about that society a hundred years down the line and designing to make sure that we’re actually designing for the things we want to see in that future.

And there are people who are doing this, there’s think tanks around there who are actually actively trying to pursue this. And it’s hard work. It is hard work to think about what it is that we want exactly. But I think the key is that when you look at that, there’s a lot of different things that come together. But at the heart of what we’ve been discussing, the Empires of AI and what AI can be doing for this future, I think is a very simple choice about what are the AI systems we are creating and where do they have human input, how do they include humans in the loop? And designing for the future we want is not gonna be easy. But I think we are in a position where if enough effort went into it, it would be possible.

 I want to come back to this point about youth being able to adapt. Right now, the AI systems have changed that dynamic. We’ve got a number of cases where we are working with experts who are working with AI agents, and we’re building agentic AI systems to work with them where they are now able to be productive in a way that is immeasurable as experts compared to what they would be. And working with the AI agents is like working with a team of interns, but much cheaper.

[00:17:35] Kate: And much more amenable to criticism.

[00:17:37] David: And much more amenable to criticism, none of the complications, so much faster. So actually, if you think about this at scale, this is the source we are hearing in the news – I hadn’t heard the term before, I think it’s NEETs,  Not in Education, Employment, or Training – young people, you know, Oxford graduates even, across the spectrum, people with no education, people with lots of education who are not finding those entry level jobs because the entry level jobs are disappearing because your expert layer are actually using AI agents to be more effective and efficient in their work, whereas in the past, they would’ve used people at entry level positions to do that.

[00:18:26] Kate: Okay, so I would go further back and say we could probably look deeper into history to see that there is precedent for what I think is a relatively recent abdication of business responsibility for training early employees, where people are expected to come out of degree programmes as early professionals and just get to work being good cogs in the system.

And I think this is still true in trades, but you have apprenticeships, you are actually learning many skills because the idea is you want to launch someone into work as a more sophisticated, trained employee. So if you are just hiring people and you don’t have the system for apprenticing them, then yeah, why would I do that? I just want AI, I’m just solving this problem. You were just a little machine doing some busy work for me anyway.

But if you really were training somebody, so they know how to manage clients, they know how to do this work, they know how to work with this other function that does very different work and kind of navigate across that, these are social skills, they’re thinking skills, they’re creativity, they’re synthesising skills, whatever they are. You can imagine that that should be very employable because I just need someone who’s enthusiastic and young and has energy and wants to learn and is a sponge. But that’s not what we’ve set up work to be for a long time.

[00:19:46] David: This has been many years in the coming, but it is really happening now. And there is this sense of, so what will this lead to? Maybe there is a lost generation now, which is gonna have a really tough time of it and I’m really sorry for those who are going through that now. But one would hope and one would expect that the damage that that will cause will almost certainly lead to society where apprenticeships are actually done more seriously, better, where there are those opportunities for young people to enter.

Right now, there is that uncertainty about, well, what are those jobs of the future going to look like, as we’ve discussed before. And so I can understand that instability is leading to this really difficult problem. And I really feel for young people coming into that. But, again, this is where I like my looking a hundred years in the future. If I look a hundred years in the future, this should sort itself out if we build the right systems, because it should mean that we get better entry path plans, better training systems to get people in.

Claiming everybody should become an entrepreneur and then making it impossible for entrepreneurs to succeed is not a good strategy. I’m not saying we shouldn’t have entrepreneurship, but if you’re gonna have entrepreneurship, what are we wanting entrepreneurship for? To build unicorns or to build good small businesses? I really think it needs to be the latter.

[00:21:09] Kate: And I think way before a hundred years, I’m going to talk about the short term, but I think these are going to be governments. You know, it’s like you look at the depression, what got through the US through the depression was New Deal era policies where you had all these work programmes, all of these different things.

I think it’s going to be government designing, apprenticeship programmes, designing different things that are getting people through probably a two year period where they start to become just competent on AI, better than the AI employees who are hireable. But you need this interim phase. Some employers will do that work, but I also see that that might be something that has to be more organised as public policy, that maybe that is the bridging.

[00:21:56] David: But let me push back on that because putting that on governments is assuming that governments work the same in different countries. There are some countries where that will be government, there are certainly some countries where it won’t be government. But my hope and my claim is that the market forces will push into this as well.

Whether you have a strong government or strong private sector, it could be government in some places, it could be private sector in others. The point is that apprenticeship programmes, apprenticeships, bringing people in, is something which is going to be. If we have a future which we are looking towards where we are seeing the jobs of the future being people with skills, then it will sort itself out.

The question is, are there enough of them and where are they, are they in the areas people want them in? That I have no idea, and these are really difficult questions.

[00:22:50] Kate: And I think so much of the focus that I often feel in an activist way ends up focusing on regulating what exists, and often I get really frustrated with the fact that a lot of activism feels very disconnected. Okay, there are problems, we need to solve them, so in your reaction to what exists, where is also the alternative vision and plan for how we get to this different future?

You can’t just say no to things and that’s enough, or just say, we’re gonna regulate this and then it’s all gonna be solved. It’s like, no, you actually have to be building things, there have to be different ideas. And so even as you’re talking, I just come back to: these aren’t AI problems, that is the gist of it, these are policy values, any number of things that are deciding this is what we incentivize as a society, this is what we reward as a society.

And then you are getting people to do very different things instead of just doing more of the same and then wondering why that system isn’t getting better or isn’t doing the things you want it to do or whatever. And so that’s where you keep referencing a hundred year timeline, but I think there are enough things that are happening out there that, it doesn’t have to be a hundred years, it could be in the next five to ten years. Really different things are conceived and force goes behind them to make different models gain ascendance, that right now just cannot, there’s nothing that’s fueling them, there’s nothing that is giving them the inputs they need to get off the ground.

And that is, I think, my biggest irritation with a lot of punditry. It’s just, it’s opinions that are reactions. There’s a lot of acceptance of certain things as inevitable, where it’s like, well this is what AI looks like, guess we’re gonna have to work with this now, instead of rolling it back further and saying, well does it have to look like this? Would it be so hard right here where we are right now to fork in new directions and be developing in parallel something that looks really different, that might, on the same timeline that big tech is telling us that this thing over here is gonna deliver impact? Why can’t we be building this thing over here with much more tangible ideas of how to change the future of work, change what things look like? But I’m just not seeing that force. 

[00:25:13] David: I think it’s really interesting, and you’re absolutely right to call me out on the a hundred year timeline. That’s nice and that’s fun for visioning exercise, but it’s not actually useful. And you are absolutely right that there are choices happening now, and that could happen now, which could put us into a, or enable things to happen over shorter timelines, five, ten years, which could be already transformative and set us up for what we want to see.

You know, I like my long term timelines because it evens out the noise, I don’t have to worry about the noise, this point about it could do it as quickly, I don’t know whether that’s true or not, it is possible that actually it might be a bit slower. But if you’re thinking about a hundred years, does it matter if it’s five or ten years slower, does that really make a difference?

It matters if you don’t get there in a hundred years, if it’s so slow, you never get there. But if you are five, ten, maybe even twenty years slower, in the scale of things from a societal perspective, if there are other benefits, might that be a price that we as society would choose to pay? Now, who is making that decision and how is that decision being made, and is there even that decision in the first place? I don’t know.

But you are right, we need to reign it in. We need to think about the here and now. What are the choices we are faced with now, and what could that lead to. And that’s what we try and do. We’re builders, you’ve been in the activist space much longer than me. It’s all very new to me, and I must admit I’m a bit more patient with it than you are because I do see the value, just coming into it now, that somebody other than me is pushing back. I don’t want to be in that space, and I’m happy that other people are.

[00:26:55] Kate: Yes. And I am not denigrating activism, I think it’s necessary, I think regulation at times is great, at times it’s a blunt instrument that is not solving the problems it’s aiming to solve. So I think that’s where I take issue, they’re so often this like, well, we just need to regulate it, well, what does that, how is it solving all these other problems? And it’s not really. So I think that’s part of what I’m taking issue with.

And also I see the reason that tech gets funded is because they have deliverables and they “ship” things, I use quotes around that. But it’s like you are holding some grand vision, but you’re not saying, okay, you’ve got the a hundred year vision, but you’re saying this year I’m gonna accomplish this, next year I’m gonna accomplish this. We need those things where it’s like, I’m not saying that tomorrow we’re going to transform what work looks like and everyone’s going to be re-skilled, but you know what? We can start doing it. And if we’re doing it at the community level instead of just like one big organisation trying to change everyone, but everyone in like different subsets, they’re working on things in their own context, and you are moving people in the right direction, things can happen much more quickly than you think.

There is stuff that starts quite slowly, and then there’s this groundswell that creates a cascade of this is how things work now, this is what we do, this is better, let’s do things this way. And those are a combination of builders and activism and community workers and nonprofits and all kinds of things working collaboratively, which is how we like to work.

I would say that is a future of work right there, that very collaborative mindset and very collaborative practice. But let’s think about what these deliverables are, not as we, but as a society. How are we building those, but also having a vision for how they start to come together, not just end up being these little, you know, spinoff lone, “oh, that works so well there, oh, there’s no plan”.

[00:28:52] David: Let me draw out one of the things you said, because I think it’s a good time to try and wrap this up, but I think you’ve drawn out what I think is the key desire. What if the future of work was more collaborative than competitive? At the moment, work, by its very nature, is competitive. You compete at the interview stage to get the job, you compete against your fellow companies to get the contracts. Competition defines work in so many contexts right now.

But that’s because it’s in some sense built around scarcity, scarcity of different things. If we are transitioning into a stage where through the innovations of AI we actually have an abundance, whatever that means, well, in an abundance environment, collaboration can and should outcompete competitiveness. So actually if there is abundance, then a collaborative approach could really be successful. I don’t know what that looks like, but I think that’s a really interesting design principle.

[00:29:57] Kate: One counter argument to that, which I agree with, but I also agree you can have cooperation and collaboration and scarcity too. I always take the Lord of the Flies example, which the way that gets told through the book is these boys on an island, they just become animals, savage animals, they’re competing, they’re cruel to each other.

But there is the real world example, and I think this has been well documented. There have been a few articles written about it, these boys who were shipwrecked somewhere in the South Pacific, and actually what emerged was a tonne of cooperation, a tonne of collaboration, they survived, everyone stayed alive.

These are actually the behaviours that most humans work toward. And when we’re in real crisis moments, those collaborative, cooperative things are what save us as humans, we are not these savage competitive animals just like destroying each other to get this one scarce resource.

Though our overlords right now might want us to have those behaviours because that scarcity mindset and having us all at each other’s throats makes us all very controllable in ways like, ’cause we’re just focused on ourselves rather than what might we accomplish together. So that’s my only argument is that I don’t think this requires abundance. I just think it requires different systems that are making people relate to each other differently.

[00:31:19] David: That’s a really fair point. And I think the idea that we’ve always recognised as an organisation is that when you are not trying to maximise profit but you need to be profitable, collaboration is your best tool. You know, this is it, it’s not necessarily the best tool to maximise, but to be in profit and to be profitable, we do best by collaborating, that’s what really helps us. That’s maybe the key thing, that it’s really about, as societies, what are we trying to do? Are we trying to maximise or are we trying to actually live and appreciate, maximise other things.

Anyway, this has been an interesting one, it got political again, and this series has been maybe more political than we’ve ever been before. 

[00:32:05] Kate: I mean, it has to be. I think what we’re getting at is the fact that most of what we are experiencing as inevitable is driven by forces that are not technological. So that’s why that keeps coming to the fore in this conversation. Okay, we’ll leave it there. Thanks so much, David.

[00:32:20] David: Thank you. I look forward to the final one.