Will AI Really Take Your Job? The Hype vs. Economic Reality | Utilizing AI Episode 40

Although we are bombarded by headlines suggesting that AI will decimate the global workforce, the likely scenario is far more nuanced. This episode of "Utilizing AI" brings Nick Patience, AI platform practice lead, and Olivier Blanchard, devices research director of The Futurum Group, to discuss the latest data on the impact of AI on the job market. Welcome to "Utilizing AI," the podcast focused on practical applications of artificial intelligence and part of the Futurum Group Podcast Network.

Every Wednesday, we explore news and use cases of the ways in which AI is transforming enterprise IT and the industries it serves. I'm your host, Stephen Foskett, president of the Tech Field Day business unit here at The Futurum Group. Before we dive into this discussion, let's meet who's on the panel today.

I'm Olivier Blanchard, a research director at The Futurum Group, and my focus is intelligent devices. So basically anything that's AI powered, that you wear on your body, that you drive, that you have in your home, et cetera. And I'm Nick Patience.

I'm the AI platforms practice lead here at Futurum, so I focus on the things that are kind of unique and peculiar to AI. And as I said, Stephen Foskett, you can find me at AI Field Day, also on the Techstrong gang most Mondays. And this Monday, we actually started an interesting conversation that I want to continue here, which is, AI is coming for your job.

Or is it? And what does that really mean? Now, most of us have seen the scare headlines that AI is going to hollow out the workforce, a quarter of all people are going to lose their jobs to AI in the next, I don't know, two years.

I actually just made those up, but I bet you thought I didn't, because that's the kind of headlines that we're seeing. And frankly, I don't know that it's entirely true. Now, Nick, you just wrote a great article, which I referenced on Monday's Techstrong gang, and I'm so glad to bring it here as well.

Talk to us a little bit about your perspective on AI in the workforce. Sure. Thanks, Stephen.

So, yeah, as an analyst, you get asked a lot of these questions all the time about jobs, so I thought I'd sort of gather it together. So every way I looked at the story is kind of the same, AI is coming for the job. So my instinct was to pull the data together and try and find out what happened.

So we looked at numbers from the US, from Europe, including the UK, the official labor stats, Fed numbers, academic studies, all those kinds of things, expecting to see some sort of start of this job apocalypse, and it just wasn't really there. And so three and a half years after ChatGPT, which is roughly where we are, there is no economy-wide job shock. Now, obviously I know that a lot of people they're obviously not saying it's already happened.

They're generally speaking, they're worried about the future. I got another one this morning, another email, a press release from a company with a vested interest, telling me that 50% of all people in Britain are worried they're going to lose their job in five years. So it happens a lot.

So, I guess, the answer is, are you going to lose your job to AI? My answer is, generally speaking, you're not, and maybe you never will. So there's various things that we looked at in the study.

There was one thing in particular, we used some of our ETR data. So ETR being the survey company that we acquired in the last few weeks, and they do these kind of quarterly surveys/surveys asking, and one of the questions they're asking is whether organizations are intending to slow down new hiring. And that number jumped, I think it went from 31% to 47%, rather than making outright headcounts, which actually didn't really move very much.

Headcount cut, that is, rather. And so more than half of the people surveyed expected hiring growth to lag behind revenue growth due to AI productivity gains as companies gradually plan to expand their output faster than their payroll. So it's obviously intent rather than something that's already happened.

But I think it kind of gives you the idea of where some of these things come from. But I still think it's more that. It's more slowing down hiring.

And then there's various things we can talk about, the various rungs of the career ladder as well. So I think the squeeze is real, but it's going to happen at the bottom rung of the ladder. Which is probably not particularly controversial.

Not the whole ladder. So I think we've seen this over and over again, with stories obviously about coders and other more entry level positions. Yeah.

So that falls in line with some of the predictions we made, Daniel Newman and I, when we wrote this book right here, "Human Machine," just before COVID. And at the time we were afraid to call it AI, so we called it smart automation. This was pre-ChatGPT, so we hadn't really made that huge leap forward into the world that we live in today.

But we knew that it was coming at some point. And we wanted to be cautious to not call something that is more smart automation artificial intelligence than artificial intelligence. But that was kind of our premise.

So we started out looking at all of the big sort of economic, and social shocks of known human history, or at least recorded human history. And the analogy that we found was sort of like the Industrial Revolution, right, which always made sense. And then it was sort of like the internet revolution.

And we wanted to look at what jobs, and what percentage of jobs, and what categories of jobs typically disappear when there's a huge technological advance. And what we found, obviously, there were a lot of jobs that disappeared, during the Industrial Revolution, for instance, the folks who used to light the gas lights in the streets every evening. There were a lot of cobbler shops and livery that took care of the horses.

All those things went away, and they were replaced by garages, they were replaced by switches or different operators. So I think it's natural and wise to predict that AI will eliminate many jobs. I think that we're starting to see that a little bit, particularly the types of jobs that can easily be automated that weren't easily automated before.

So just like we saw in manufacturing, somebody who put a peg in a hole over and over and over again was replaced by a machine. And I think that these simple, repeatable, predictable tasks that are done by humans today, that can be done more cheaply by a machine by virtue of the fact that we live in a capitalist society, will see the same trajectory. And so a lot of very low level or low skill jobs, and I want to qualify that because I think that a lot of things, a lot of jobs that we refer to as low skilled are actually not low skilled at all.

They're just low pay. But something that is very repetitive and predictable will be replaced. And I think that the closer a job looks like a task as opposed to a series of complex interconnected tasks, the more easily replaceable or at least the most easily automated it becomes.

For most of us, that's not the case, right? I won't talk about the analyst world because that's not the point of reference that everybody who's watching this understands. But if you think about your job and think about all the things that you have to juggle every day, all the problem-solving, all of the catering to people's incompetence or other flaws that you have to just kind of cover for and fix and all the problems that we have to, all this churn that's created around us all day that we have to deal with, it's very difficult for a machine or an AI to deal with all this and have the competence and the experience and the mental notes and the context that we as human beings have.

And the more professional the context, the more difficult it is to just sort of automate any of this with any kind of real true positive outcome or without creating more problems than you're solving. And so I think it's incredibly complex for us, no matter how good the AI gets, to replace jobs that deal with intangibles all day long, that are not that predictable, not that organized, not that square. And so I'm with you, Stephen, and also with you, Nick, in the sense that I think that AI can do a lot of things for us.

I think that AI as a tool or as a toolkit can help us become more productive. It can help us move faster. It can help us do things better, make fewer mistakes.

There are a lot of things that it can help us do, but ultimately It works best for the most part when it's working around a human who knows what they're doing as opposed to this concept that, well, AI, because it can do certain things, will replace a large percentage of jobs anytime soon, if ever. I don't see that happening in the next five years certainly, and probably not in the next 10 either. Yeah.

I want to jump in with a metaphor here. Now, for what it's worth, I actually majored in the impact of technology on society when I was in college, and I find myself studying that even now. When I'm not working at Futurum, I'm actually a historian of watchmaking, and one of the things that I have written about on that side of things is the rise of the American mass production system, and I think it's relevant here.

So let me give you guys a little context and see if this metaphor is appropriate to, well, I guess our listeners, which are in the enterprise technology, software development kind of world. So a lot of people mistakenly believe, including the Americans at the time, that American industrialization happened because of things like the assembly line, but it was actually sort of the reverse of that. Those things were developed because of the inputs that the American manufacturers had, which is to say no skills.

So essentially, what happened when American factories started to appear is they were trying to hire people into those factories, but they had a lot of people, but no skills. And so the successful development of American manufacturing in the second half of the 19th century happened when they developed production lines that could bring in workers with no skills, literally zero skills, literally a guy who just came in off the field or a woman who just came in from the home and who could take a spot on an assembly line without having to have gone through an apprenticeship and been trained and so on. And that worked phenomenally well, from the Springfield rifle to the Waltham Watch Company to Henry Ford, in building this massive scale of production.

But unfortunately, it also came back to bite American industry. And one of the reasons that we don't have American watches anymore, and one of the reasons that American automobile production fell in the '70s and so on, is because the Americans didn't invest in skills. They kept investing in no skills and kept investing in machinery rather than in people.

Whereas if you look at countries that were more successful, like the Swiss watch industry, but also, for example, like the Japanese automotive industry, they learned ironically from Americans like W. Edwards Deming that investing in better and better workers was the way to improve the product in the long run. And so in the case of the Swiss, they basically continued this incredible focus on apprenticeship, and they were able to come out dominant in this field in the 20th century.

Now, having an industry that didn't get destroyed by war also helped, but they were able to do really well because they had this incredible reservoir of skills, and those same skilled people were able to bring that industry back in modern times. And it's the same when you look at a lot of other industries where it's, I don't want to say easy, but it is fairly straightforward to develop massive scale at the expense of future success. It's much harder to develop future success because it means investing.

And if you can see my metaphor, when I'm thinking about software development, my concern is more that we may hollow out skilled workers at the expense of thinking that lines of code or pages of text or ridiculous images to illustrate blog posts is a sign of success rather than seeing that as a sign of failure. So Nick, I see you nodding a little bit. Maybe that works for you?

What do you think? Yeah. So are you also saying, though, that you'll hollow out the skills because the senior developers, in this case coders, have stopped being developers or coders and become managers?

Is that the concern? Well, that's actually another thing that I'd like to talk about. Okay.

First, I was starting with this metaphor that essentially, as an industry, I think that we need to really focus on people development and- Yeah ... and creating a system that improves the overall expertise rather than just creating a system that writes more code. Yeah, I'd say that's...

Yeah, I wholeheartedly agree with that. I guess it's how you go about it, isn't it? And obviously there are, we all know that there were and there are thousands of computer science graduates being churned out every year, if we're sticking with the coding as the job, who then can't find employment.

So I guess they've been skilled up, but I guess in that case, maybe they've been skilled up in the wrong skills to a certain extent. I'm not entirely sure. It's interesting you mentioned Ford because there is that example in the report I wrote about how we try and look at these various things, and there was this Jevons paradox, and there's Engels and various other things.

And one of the examples was that Ford actually had to rehire some of its veteran engineers after the AI that it tried to deploy to do part of their job fell short. So it's still happening to a certain degree. So I guess in that case, they thought they didn't need these more senior people and were trying to replace them at a considerably lower cost, but it didn't actually pan out.

Yeah. Sounds like the conversation we're having is quality over quantity, right? So with your watchmaking example, Steven, Swiss movements and Swiss watches, it's more luxury, it's premium.

You're not just buying a thing on your wrist that looks good. You're buying something very specific, a very specific set of values. Whereas if you buy, and I'm a big fan of Casio, but if you buy a Casio at Walmart, it's because it's $25 and you need a little electronic on your wrist.

You don't really care, right? They're all the same. And it just does the thing, and then when it dies, you throw it away or you replace the battery.

It's not a big deal, and you don't need to invest in it. And I feel like somewhere that's the conversation that we're having, that some companies will invest in these quality skills, right? And the importance of that depends on the industry, right?

If you're hiring engineers to design cars versus using AI to design cars, you're going to have very different results. And you're going to appeal to very different parts of the market, and I think that your price point and your business model is also going to be very different. And fundamentally, I think that AI feels to me, for the most part, like the quantity part of the equation.

And humans, skilled humans, talented humans, people with experience who can do, or expert craftsmen and craftswomen, craftspeople in their own areas or in a cluster of areas that can just sort of mix things together and create something new are the quality. And companies that find a way of balancing both or really zero in on the one part of that binary that fits their model best are going to do really well. And companies that don't understand this dynamic and try to mesh the two together or try to get quality out of the quantity piece of the equation, and people who try to get quality out of the quantity piece of the equation are going to find themselves in some difficulty in the next few years.

Yeah, and I don't want to be misunderstood. I'm not saying that your Casio is terrible. In fact, quite to the opposite.

Yeah. The problem is that if you treat your people like machines, they don't develop the next generation of product. And in Japan, they absolutely did develop the next generation of product, and that's why your Casio is great and still exists.

In the US, the workers were machines, and they were not- able to develop into the next leaders and the next inventors and the next designers, and their children didn't go into the field, and that just didn't become the progression of the industry as a whole. And when it comes to AI, I think that this is a concern as well. We've heard a lot from educators about how AI is being used by students.

" And that's a bigger question. What does it do if we begin to use tools so much and to rely on tools so much that they become a replacement for our own capabilities? One of the things I want to bring up that I brought up as well on the Techstrong gang was the aviation industry and the medical industry.

So in the case of the aviation industry, there was a lot of concern after the 2009 crash of Air France 447 because the extremely senior pilots had come to rely too much on autopilot and no longer understood how to deal with crisis situations. And that was a whole industry, a whole area of scientific investigation. Similarly, there's been investigations into the effect of AI on doctors.

When they come to rely too much on the AI to diagnose, they quickly lose the ability themselves to diagnose. And I'm concerned what happens in software development. If we have Claude Code doing not just our coding, but also our QA and testing and documentation, then do we truly have those skills anymore?

And what do you think of this, Nick? Yeah, I think it sounds plausible. I think the aviation industry is...

The autopilot's been around since what? '70s, isn't it? So it's incredibly ingrained in how planes are flown.

So there's been a very long time for it to have taken root and been the kind of acceptable way, and so that does make a lot of sense. We're at quite a different stage with AI. Obviously, we're at a very early stages.

And so I think timing is everything. There's that classic example that every quotes about bank tellers, wasn't it? And it was turned into a great example by the economist whose name I put in a report and I can't remember exactly.

But then we sort of saw in the early 2000s of James Besson, that's it, made it as a kind of rebuttal to a kind of automation panic saying there are more tellers, and obviously banks started opening a lot in... I was living in the US in the '90s and 2000s. It was amazing how many banks were opening.

But then mobile banking came along, and then that just stripped out those jobs. So I used to have a great slide of, I think it was Royal Bank of Scotland, which for a very brief period of time was the biggest bank in the world. " Because mobile banking became so sophisticated, it could do all the things tellers could do and a whole load of other things.

So it was a timing thing. It wasn't wrong. It just got the timing completely wrong.

So this is why I wrote the thing was partly because we're three and a half years past ChatGPT, and we keep hearing these things. So it doesn't mean that there will be no job losses, and I think we're all agreeing that there will, and there'll be task automation all over the place. It's a question of can you get the timing right?

As an analyst, and I'm not an economist, but if you were, as an economist. And so I think the lesson isn't the automation anxiety isn't always wrong or isn't always right. It's about the timing and the kind of the completeness of how you substitute things.

So I'm always wary of both doom mongers and debunkers kind of trying to declare victory a bit too early. Yeah. Since my name is Olivia Blanchard, I share a name with a famous economist of the same name who we get confused with all the time.

Yeah. But I'm not an economist, but I do understand the basics of if a certain percentage of the jobs disappear, living in an economy where wages essentially drive the economy. If we lose too many jobs and we have too many people who no longer earn a living and no longer earn wages, what ultimately are these companies selling products to or services to?

Ultimately, there has to be consumers on the end of this that pay for these services. The B2C sort of is the foundation of the B2B that sits on top of it. And so if we replace all the workers or a large proportion of the workers with AI, and they no longer earn and they no longer can spend, it doesn't work.

And so, I'll just add this into our conversation, that there's a bigger economic issue here than just whether or not AI can or should or will replace most of our jobs. Fundamentally, the capitalism doesn't work if we don't have wages and workers. And somebody needs to explain to me what the next thing is other than some kind of techno-foetalism if we continue to move in the direction of all jobs will be eliminated by AI.

It just doesn't make any sense to me. And if you look at, again, the history of science and technology and on society from the famous Luddites through to the widescale anarchist activity that opposed industrialization, especially in Europe, to socialists after World War II, all of them had that same concern. Essentially, it's all well and good to automate us out of a job and so that you can make more money, but then who's going to buy your stuff?

You have to have society still. Yeah. But on the optimistic side, what has generally happened is new industries have emerged, new jobs have emerged.

My father certainly never could have imagined what I do for a living, and probably yours as well. And I bet his father could never have imagined what he did for a living and vice, and on and on down the road. Society and technology do advance, and it's not all doom and gloom.

I think that that's one of the things that people may be thinking. " But is it? What's your response to this doom and gloom scenario?

No, I don't subscribe to it. I'm cautiously optimistic. I think it's the kind of Jevons paradox stuff about if something valuable gets cheaper and better, we want more of it.

And so I know I was listening to, and I've included some of it in the report, Aaron Levie, the CEO of Box, who's a great commentator on the tech industry. He was saying about how he calls himself Jevons pilled, but he's saying he sees engineers in his organization becoming more efficient, and they get more work done, and then they want to do more work, and they need more people. And I guess history is a graveyard of confident job apocalypse predictions.

The famous various ones we've mentioned, the tellers. Jeff Hinton in 2016, "Do not train to be radiologists. " There are now more radiologists, and the Mayo Clinic was a really good example of they have more radiologists, but they also have a team of data scientists, which didn't exist.

There's a job that didn't exist 30 years ago. Digital video editors didn't exist. All sorts of jobs didn't exist and say like ours didn't.

So I think if the squeeze on the entry-level rung of the ladder starts to move up seniority, then we have a genuine problem. But there's not a, what's it called, the lump of labor fallacy. There's not a fixed amount of work to go around.

That's just not really how things have developed with humans. So I think there will be more jobs. We're also facing, it's a different podcast entirely, but population crisis in various countries.

So in some countries, they're actually going to be very short of people. So there's going to be different speeds of automation in different countries. And so, yeah, I kind of tend towards the positive on this, and if the labor stats start to show seniority, seeing more senior jobs being automated away, then we may have to revise our opinion.

But for now, I don't think that's necessarily the case. Yeah, I'm neither optimistic nor pessimistic. I might lean a little bit more pessimistic right now, and a lot of reasons for that we don't have time to get into.

But part of it for me is also that I still don't see real scalable demand for a lot of the AI that we're seeing right now. I feel like it's more of a supply trying to create demand rather than demand for supply. And even the economics are still very upside-down with that.

And I'm seeing a lot more negativity in the user space around obviously data centers being built in their backyards and flock cameras and surveillance tech and all these things with, on top of that, this specter of is this coming from my job? Am I going to lose my job to this on top of everything else? And so we haven't really seen that many compelling use cases for the average person outside of very specific fields like engineering, like medicine.

And so the case still needs to be made to users for the need for AI and how AI isn't going to be this apocalyptic transformative thing that's going to eliminate everybody's wealth and freedom and just make a handful of people very, very rich, just as the climate crisis just kind of explodes. And so it's hard not to be a little bit pessimistic because that is definitely one possible trajectory for this. It's just I don't think I've seen a concerted effort or even a credible one to give us other alternatives outside of science fiction currently, and I think that's something that's sorely needed.

Even from an ROI standpoint, the case studies, even the ones that we're looking at, are few and far between when we look at the cost of implementing or integrating AI with actual really good, solid, repeatable metrics even in the enterprise. There are still too few of those. Of course, we could play a part in changing that.

I think we have a responsibility as analysts to do that. But still, as Nick said, very early days, and I think that we need to have a more unified vision of what AI could be in service of humanity as opposed to being in service of a half a dozen companies. I'll just add one thing to what you're saying, Olivier.

But if we don't have use cases that are actually viable outside of narrow domains, then this thing won't take off, and therefore people will still have their jobs. So it's kind of - Yeah ... so we won't need the technology.

So if this whole AI thing, which I've been looking at for a quarter of a century, doesn't take over the world, then yeah, we'll be fine, too. So- Yeah, I think the pendulum might swing a little hard both ways. I think that we're going to probably...

I'm already seeing some companies sort of blame AI for layoffs, even if AI- Oh ... has nothing to do with it. So we're seeing this intent that, well, maybe AI will save me a lot of money, and I don't need all these people.

But then ultimately, what ends up happening is like, "Oh, no, we're not ready yet. " And I think we're going to go through cycles of this the next few years that should be actually fairly predictable. Yeah, and I think that it is a lot more nuanced maybe than a lot of people say.

" That I don't think really helps the situation very much. " Maybe take that with a grain of salt. That is marketing.

And frankly, it's effective marketing because essentially it has convinced everybody to get on board and try to adopt this. But that being said, that doesn't mean that it's complete nonsense. I think that what we've seen as well is that when transformative technologies come, they do disrupt the job and labor force to a great extent.

And frankly, we also have a pretty poor track record of job retraining, especially for older veteran workers who no longer have their traditional craft to perform. That being said, in many cases, those folks are able to move out of the workforce, retire, and allow the young people to step in and do something different. So I as well am not entirely pessimistic about this.

I think that we will find a way as a society to adapt to this, and I think that we will move forward. But one thing I will say is that, if history teaches us any lessons, I think that government needs to make sure that as we're doing this, that we don't allow the situation that you described, Olivier, where just a few people control everything because that artificially limits the ability of society to adapt. I think we need to let thousands of new people experiment and try new things and so on, and not get squashed by the incumbents who are leveraging this technology to secure their place in the world.

Just an incredible discussion as always. I have to say, I have so much enjoyed Olivier and Nick, having you all on "Utilizing AI," and just an incredible conversation. Thank you so much for joining us.

Before we go, I'll give each of you a chance. Tell us where we can connect with you and continue this conversation. Olivier?

Well, you can connect with me obviously on the Futurum website, where most of my research and articles and thoughts are published, but also on X, occasionally on LinkedIn. You can always send me a message there. I don't publish a lot there, but I'm present and attentive.

com. I am more present on LinkedIn. My "Patience on AI" newsletter, which comes out every two weeks, was where I first started kind of ruminating on this topic before I then wrote the report, which is also available on Futurum's website.

I'm on X occasionally as well. But yeah, you can find me around. And I will say that "Patience on AI" is absolutely indispensable reading.

I look forward to it every couple of weeks. Olivier, you should do that, too. That'd be great.

I should. That's true. And as for me, of course, yeah, you'll find me on LinkedIn.

Also, I am on some of the alternative sites, Bluesky and yes, Mastodon. I run my own Mastodon instance, so connect with me over there. Thank you very much for listening to this episode of the "Utilizing AI" podcast.

If you enjoyed today's discussion, please do subscribe on YouTube or in your favorite podcast application, and consider giving us a rating and a review. This podcast was brought to you by the analysts and experts from the Futurum Group, where insights meet AI. ai, the "Utilizing AI" YouTube channel, or the Techstrong TV app.

Thanks for listening, and we'll see you next week.