Are jobs getting better?

In the UK we have 1 in 4 workers who are in what I would consider to be a poor quality job.

In London, that's even higher. It's 1 in 3 workers. So these are workers who have poor employment conditions, often low wages, precarious contracts, unstable jobs, work too many hours or work too few hours and are looking for more work.

Host: That’s Kirsten Sehnbruch from LSE’s International Inequalities Institute. For many, the world of work is defined by instability, poor contracts, and rising insecurity. On the other hand, technological shifts are offering unprecedented gains: AI promises a massive productivity uplift, removes drudgery, and stimulates creativity. We are seeing models of employment which promise the ability to work from anywhere, providing geographic flexibility and making location no longer a constraint for individuals choosing where to live. And overall, on average, people are working slightly less.

Welcome to LSE iQ, the podcast where we ask social scientists and other experts to answer one intelligent question. I’m Maayan Arad from the iQ team where we work with academics to bring you their latest research and ideas, and talk to people affected by the issues we explore.

In this episode, I ask 'Are jobs getting better? And, if so, for whom?'

I learn why we need to focus on job quality, not just quantity and discover how AI is reshaping creativity and productivity. I also find out if a new model of work could finally stop the 'brain drain' from smaller towns.

We’ll return to Kirsten Sehnbruch’s warnings about the precarious nature of the workforce later. But first, let’s hear about the geography of work. Raj Choudhury is a Professor of Organisational Behaviour in LSE’s Department of Management. He argues that the 'Working From Anywhere' model isn't just a pandemic hangover. It's a unique opportunity to solve deep-rooted economic and societal problems.

Raj

...I think the geography of work is changing quite rapidly. And it was changing prior to the pandemic.

But the pandemic acted as a sort of like event that precipitated many other companies to embrace remote and distributed work practices. ...the reason I was excited about studying work from anywhere, ... is it gives employees and individuals a lot of freedom to choose where to live, while giving companies the ability to hire from more places. So I see work from anywhere as a win-win. It's good for individual workers and it's good for companies.

Bridge: Raj explains to me the difference between the concept of working from anywhere and the more familiar idea of working from home.

RAJ

...So work from anywhere is a work arrangement where the worker gets to choose where to live. And it may not be complete freedom, but some freedom...

In some cases, it might be the city, the state, and even the country… So if I'm a young worker, housing costs in a city like London or any large city in the world is prohibitive or not. but if my company allows me to work from anywhere I can go and live in a smaller town.

…I could go and live in Leeds ... Liverpool or Bristol, ... the housing costs are much lower. So it becomes more affordable. And then I can afford a bigger house. Work from anywhere allows individuals to live closer to their parents, and many people have caring responsibilities. So...that's a very important priority for many, many individuals.

The third... So for many couples, getting two jobs in the same place is a struggle. So the dual career problem and there's both research and I have anecdotes of women losing out on career progression because of the location.

So if you're a woman living in London and the company is giving a promotion but asks you to relocate to New York, maybe your partner or your kids don't want to move. But if your company gives you work from anywhere as a flexibility, then you don't have to move for the promotion.

And what about employers? ... some of them might be concerned that ... productivity...might be impacted when their employees are working from other places. That they have less contact ...than you would have in an office setting. What is your perspective on that?

RAJ

So the biggest benefit for employers if they embrace work from anywhere is they can hire from anywhere. So instead of getting all the talent from one single location, now your labor market expands. So you can now hire from not only London, but you can hire from the rest of the UK. You might be able to hire from other European countries, maybe North Africa, Latin America, even Asia.

So the world becomes your labor market. That's the biggest benefit. And my research has shown that work from anywhere could lead to productivity increases .... But let me also quickly add something here. I'm a strong believer of work from anywhere being beneficial to both companies and workers, but I'm also a strong believer of in-person.

And when I say work from anywhere, it doesn't mean that we never meet in person… I actually think of different forms of hybrid work that allow for both work from anywhere and in-person time.

Bridge: But getting that balance right can be difficult. We’ve all experienced the frustration of commuting into an office only to spend the day sitting on video calls that we could have done from home. For Raj, this defeats the purpose. He argues that if we are going to ask people working from anywhere to come in, we need to completely rethink what that time is actually for.

RAJ

...The first design principle is we should not be working on individual tasks. Because right now, what's happening is many people are going to the office or wherever, and then they're wearing headphones and getting on zoom, or they're closing the doors and doing coding, or ... writing their reports.

And I see no sense in that, because those individual tasks can be done from anywhere. So when we are in person with the team...those days should be scheduled as only team based activities. So it should be focused on mentoring. It should be focused on brainstorming. It should be focused on collaboration exercises. And most importantly, it should be focused on building trust.

MAAYAN

Often when we think about...working from anywhere, it's mostly big tech companies or the private sector... that comes to mind... Is this something that can also work for the public sector?

...work from anywhere is important for all companies, but especially for public sector, because the public sector cannot compete with the private sector on wages or salaries. So what the public sector can offer is flexibility. So I think work from anywhere is actually more important in the public sector compared to the private sector.

... So earlier, the criticism of work from anywhere or work flexibility was that it's impossible to do flexible work or work from anywhere if you are a doctor, or if you're a nurse or you're a factory worker.

Now, because of digital twins, it's possible ... even in these settings. So I think... work from anywhere is going to now become more popular in sectors that was not possible even like five years back.

Bridge: Raj describes these ‘Digital Twins’ as a system where sensors and AI create a real-time virtual replica of a physical site like a hospital or factory allowing professionals to monitor and manage them from anywhere. For Raj, this isn’t just about efficiency but also a mechanism to reverse the brain drain that has diverted talent from smaller communities for generations.

But while "work from anywhere" changes where we are, a different force is fundamentally reshaping what we are capable of doing. Artificial Intelligence is the elephant in the room. And according to Fabien Curto Millet, the Chief Economist at Google, we shouldn't underestimate the historic scale of this moment.

Section 3: AI (productivity gains with Fabien)

FABIEN

...this is the most profound shift that I've come across in my professional career as an economist...We haven't quite seen something like this. And firstly, it's something quite rare that we were observing.

...AI is general purpose technology. It joins a very select club of technologies. We've got, you know, a steam engine, electrification, personal computing, and now you have AI. And these are technologies that transform the economies really profoundly. So it sort of meets all the characteristics of a GPT.

...I think and this is quite special, that AI is also what you might call an invention of a method of invention. So the late Nicholas Crafts, you know, had this expression and was already pondering where AI might be, and so what we mean here is that AI is not just a single breakthrough, but a breakthrough in breakthroughs.

And so what you see, especially when you apply AI to science, is that it accelerates the pace of scientific research. And the best example of that is, you know, it's a homemade example. So, AlphaFold, which is a system released by...Google DeepMind that won the Nobel Prize to a couple of my colleagues, Demis Hassabis and John Jumper, in 2024.

And essentially it predicts the... 3D structure of 200 million proteins and that's now released, you know, openly for millions of researchers across more than 190 countries across the world to using their research. So, I mean, the use of it is formidable. But to give you a sense of how transformative this is, to predict or to understand the structure of a single protein that could cost you hundreds of thousands of dollars, and potentially years in the lab.

And suddenly, thanks to AI, in one fell swoop, you've got 200 million of these that are out. And so now you've got people working on malaria, vaccines, cancer treatments. I mean, it's formidable, the unlock, but it has facilitated. So that's why this is, you know, more profound, I think, than even other GPT, than we have seen.
FABIEN on productivity and creativity uplift 00:02:45:15 - 00:03:02:24

And, you know, it's coming not a moment too soon in terms of the challenges we have as societies, in terms of the adverse demographics we face. ... so it really has, you know...the potential for a massive productivity uplift.

Bridge: Recent studies have shown tangible productivity gains ranging from 14% to 40% across professions like customer support, coding, and professional writing. For Fabien, these micro-improvements are already adding up to significant time back for the average worker.

FAB

...we found that workers on average, you know, could make time savings of 122 hours in a year. So the productivity evidence is tantalizing. It's tangible at the micro-level, the question is, when is this going to translate to the macro level? And that's a whole other ball game. It's early innings right now.

There's higher creativity, and, you know, higher quality of output in the work. And so these dimensions of well-being at work, are also things that we find, we found in our own research. ...Eric, bring you some paper on call centers, which is a very dense environment, right? I mean, customers are sometimes not happy when they're talking to customer service agents, but Eric and his Stanford colleagues found that there was higher retention of agents who used AI, because the conversations were less hostile.

There was less escalation to managers.

MAAYAN

What do you think people get wrong when talking about AI and jobs? And, you know, which very common claim doesn't really match the evidence that we have so far.

FABIEN

...there's a lot of sort of, you know, discussion out there around sort of, AI impacting jobs as a whole. And, that's not how technology works.

So, you know, the way technology works from an economic perspective is it automates individual tasks. On the job, it is really a collection of tasks. And so a job only gets fully automated away when a significant proportion of the underlying tasks are automated, which is a fairly rare event. So Jean Besson, looked at the 1950 census, 270 occupations in the US and found, I think he published in 2016, but only one had been automated away, which was elevator operator.4

... So occupations kind of do sort of better and worse.

...And on the whole, technologies are a net creator of jobs, it routinely does.

The other part is around the idea that AI is a sort of white collar affair. But it only impacts the cognitive realm. And of course it does. It very much impacts the cognitive realm. But I think people are under appreciating how transformative it is going to be for blue collar occupations as well.

So in the US, we have a shortage of 100,000 electricians right now. Similarly, I mean, it's slightly lower, but, large scale shortages of welders, Hvac engineers, and a lot of that is driven by, you know, the AI, boom, essentially where, investment in data centers and all of that build out is, requiring, a lot more people in those, you know, traditional trades, who are thriving.

So that's doing very, very well, even short of their occupations being modified day to day. But I think those occupations will be modified, day to day, because if you take a step back, what AI does essentially is it radically reduces the price of knowledge and analysis in the economy or reduces the cost.

And so knowledge and analysis is present throughout the economy. And of course it’s present in blue collar occupations.

Section 4: Low quality employment

Bridge: Democratising access to knowledge is one part of the equation for a better labour market. The other part is ensuring the jobs themselves are sustainable for the people doing them. Kirsten Sehnbruch’s research looks past the technology to focus on the human experience of the employment contract. She estimates that one in four workers in the UK are currently in 'poor quality employment'.

KIRSTEN

So at the moment...the only way we compare workers is their wages per hour. But say for example, you've got, you know, John who's on a zero hour contract working in the retail industry at the minimum wage. You have, Sophia, who is working on a platform with very irregular wages, but also basically making the minimum wage, but never knows how many hours she's going to be working.

And you have somebody who is, let's say also a minimum wage earner and has a stable job, has been in the same job over ten years, has access to training, job security, etc. these three workers are all minimum wage earners. And so saying, okay, they're low income earners, that isn't going to help us. So one of those workers is going to be the most deprived, and one of them is going to be the least deprived.

And our public policy should be oriented towards the one who is the most deprived. Taking into account also their personal circumstances and the ability that they have to convert their earnings potential into employment capability. So somebody who is looking after a small child, for example, and is restricted in terms of working hours, that person will need very, very different support to somebody who hasn't got any caring responsibilities.

Bridge: We often assume that skills automatically lead to good jobs, but for many workers, especially women, the need to care for children or elderly relatives forces them to make a difficult trade-off. They have to accept jobs well below their qualification level simply because they need the hours to work around their lives.

Kirsten

...they adapt to the fact that they have caring responsibilities and therefore they often have short term contracts, permanent, part time contracts, flexible contracts, etc..

...the problem with flexible jobs is not the fact that they're flexible, it's the fact that they often come and are associated with other employment conditions that are not very good.

Bridge: This is a crucial distinction. "Flexibility" sounds like a perk, a way to balance life and work. But if that flexibility comes with low pay and zero security, it creates a trap. And Kirsten argues this isn't just a problem for the individual worker trying to pay their bills. When you have millions of people trapped in low-quality work, the cost eventually lands on the taxpayer's desk.

Kirsten

...there is a tipping point in this model where if you have too many workers who are in poor quality jobs, then that requires an awful lot more resources on the part of the state to sustain them. So for example, in the UK that would be universal credit. And at the same time, people in precarious jobs pay less into welfare benefits, either because they're working part time or because they're working flexibly, or they have interruptions in their employment trajectories.

And so at the end of the day, you're earning less from those jobs as a government, but you're also paying out more to sustain them.

What daily problems do people describe most often when they talk about their job? ... Were there any stories in particular that stayed with you?

...So what we often hear is how difficult life is when you have small children, because you have to balance your work situation with childcare. And if you don't have grandparents close by who are on hand to help, that can be extremely difficult.

And so that's a very sort of typical middle class problem of a particular age group. When that gets really difficult is when it's a low income situation. So typically like a single mother, somebody I know, who has three children, nobody around to help with caring responsibilities. And you will either have a shortfall in terms of how the children are cared for, or you will have a shortfall in terms of employment.

..the level of stress that that causes in families is, in my view, in the long term unsustainable. And when you talk to people, I mean the stories are sometimes heartbreaking.

And what we... end up with very often is people who, especially mothers stepping out of the labor market so that they can take care of children because they can't do it both, or because their earnings don't cover childcare costs.

And the problem with that, of course, is that the minute you stop working, your skills atrophy. It becomes much, much harder to reinsert yourself in the labor market.

Bridge: And this creates a catch-22. You step out of the workforce because the childcare costs are too high, and often the only way back in is to retrain. But for many people with a mortgage or a family to feed, taking two years off to study is a luxury they simply cannot afford.

KISTEN

... if you are older and have a family, you often can't stop working in a job that's of poor quality because you can't afford to retrain.

You can't just say, okay, for two years, I'm going to be without earnings. So other countries, the case I know best is Germany, where you can apply for not only the new training. Often that comes in the form of apprenticeships or sustained courses over, you know, 1 or 2 years, etc. and if it's an apprenticeship, you will be getting a small wage from that employer and then the government will compensate, taking into account the earnings that you had before.

They will make up the difference between an apprenticeship wage and your previous wage, up to an amount of like 80%, or other forms of funding, education like that. So I think the biggest question that we need to think about in the UK is once people have entered the labor market, how are they developing going forwards and who is funding that further development?

Kirsten

So looking at just technologies isn't enough. ... is universal credit going to be enough to help people into new jobs? Or do we need a, you know, training for life program where a university or apprenticeship is just the first training that you, you get or that is funded?

... We don't really have an infrastructure in place... At the moment. It's all quite ad hoc. So that's, I think, the biggest, policy gap that we need to fill.

Bridge: Kirsten argues our current safety net is too reactive. It waits for you to crash before it steps in. In an ideal world, she would like to see Britain rolling out an "Employment Insurance"

KIRSTEN

...So not unemployment, but employment where you have in one system, a sort of three tiered level of support, one which helps companies and advises companies on how to adapt. The second to help workers adapt to new skills, training, qualifications, etc..

And the third would be if all else fails, then a decent system of unemployment insurance that can genuinely help workers over a longer period of time, not just be unemployed and survive, but also, find new employment through new training and technology. So what we have at the moment is really a system that pushes you into the next available job, regardless of how good that job is, how many hours it is, as long as they get you over 30 hours.

____________________

Bridge: Kirsten’s arguing for a safety net that helps workers retrain and adapt to new technologies rather than grabbing the first low-quality job available. But in recent months there’s been a growing fear that one demographic is particularly vulnerable. If AI can write basic code, draft emails, and process data, what happens to the entry-level employees who used to do those tasks to learn the ropes? Fabien Curto Millet points to a much-discussed study that claims these young workers are the "canaries in the coal mine", the first signal that human labor is being replaced.

FABIEN 00:12:41:11 - 00:12:59:22

...So the best paper on this topic, even though I disagree radically with its conclusions, as I'll say in a moment, but it's, again by Eric Brynjolfsson at Stanford. And him and his co-authors believe they have identified, you know, the proverbial cannery.

So they have data from, you know, millions... of workers essentially via ADP, which is a private payroll processor, in the US. And the headline finding of the paper is that there is a 13% relative decline in employment, for early career workers ages 22 to 25, in highly AI exposed occupations.

So it's a little bit of a mouthful, but that's kind of a finding. And essentially, if you look at their paper, they find that, you know, right on cue, you know, ChatGPT launches in November 2022. And from that point onwards, the fate of these sort of young workers in highly exposed industries starts diverging from the rest of, you know, the economy.

So that's what they find, you know, great paper. But the interpretation for me is, you know, totally not correct in terms of pinpointing I, and the reason is these guys are looking at employment numbers. Whereas if you look at job postings, job postings actually started to shift six months ahead of the curve, ahead of ChatGPT coming out.

And so this is April 2022. What was happening then? The Federal Reserve started the sort of, largest, interest rate hike in 40 years. So it was, of course, trying to control inflation, very massive sort of macro event. And from that point onwards, you find sort of a finding in this, this paper, but absolutely linked, and timed with the macro event.

Why is it that highly AI exposed occupations seem to be more affected? Because we believe that those are also more sensitive to interest rates. So actually, what we're picking up in this paper is interest rate sensitivity. And, again, I'm referencing this paper from Denmark and Vestergaard. Phenomenal paper. They look at the same problem for early career workers as the Stanford researchers and find, you know, but there is no impact of AI.

And they have very, very rich data there. So my read of the literature up to this point is that, you know, AI is not, sort of the culprit of the current situation, but it is a low, high or low fire labor market, certainly in the US. And that affects mostly people who don't currently have work, which is predominantly young people.

Bridge: Our guests (Fabien) pointed out that these are still early days and we need more data to better understand this phenomena but if the current slump is driven by interest rates rather than automation, the future might not be so bleak. In fact, Fabien reminds us that historically, when a new technology arrives, it’s usually the younger workers who win.

...young people are typically the ones who make best use of technology. We saw that in previous waves.

It was the sort of older generations who where we saw evidence of, you know, like excessive retirement or higher retirement than expected, who couldn't quite keep up with the young people thrived that they believe.

Maayan (record): What advice would you give to young people entering the labour market?

FAB

what I tell young people is become masters of AI. Make sure that, you know, you leverage it to your benefit, and to develop experience, develop judgment, so that you are leaning into those human qualities that are irreplaceable.

Bridge: Leaning into those human qualities is how we stay relevant. But if we zoom out, new technologies come with the promise of something beyond job security. They offer a new way of living. I asked Raj Chourdry what a "better job" truly looks like for individuals and how it helps them live their life to the fullest. For him, it’s about finally solving a dilemma that has forced workers to compromise for generations.

_____________________

RAJ

...for far too long geography has been a problem for most people. For many people, I wouldn't say most people. So you have to leave your parents and migrate, or you have to live in a very high cost city and somehow convince yourself that that's the right thing to do.

The quality of life would be much better if people are allowed to choose where to live. So if I like the mountains, I can go and live in the mountains. If I like the ocean, I can live close to the ocean. If I have caring responsibilities, I go live close to my parents. If I'm a young person with 2 or 3 kids, I can go and live in a smaller town and get a big house.

And if for childcare. So I think we can turn geography into our friend, not our foe.

Bridge: I also asked the central question to both Fabien and Kirsten: looking at the big picture, are jobs actually getting better?

FABIEN

I think it's getting better. And that sort of, relies I mean, all links back to the discussion we had earlier around, you know, more empowerment of people being able to do more expert work, less drudgery. That's the main one, and more creativity or ability to sort of tap into creativity.

00:32:29:02 - 00:32:46:19

...in many sectors, we're not running out of work. But we're running out of workers...

00:32:46:19 - 00:33:09:12

And we could really use a productivity boost. So, I'm looking forward to that, to AI rolling out and to it rolling out responsibly because it's of course, going to require adaptation, not at the level of cohorts, which is, but most comfortable means of adaptation of a labor market. But at the level of individuals, because it's happening within our lifetimes very, very fast.

And all of us are going to have to, to tool up.

KIRSTEN

...So if you look at it from a public policy perspective, actually what we always tend to find is that wages are going up in part because we increase minimum wages so that, decreases the level of earnings deprivation, especially if minimum wage is increased by a little bit more than inflation.

This doesn't do much about the overall distribution of earnings because higher earnings increase by much, much more than minimum wages. But in general we'll find that earnings deprivation is either stable or decreases a little bit. And the other thing that has improved significantly in the UK is access to pensions, as employer pensions have been rolled out.

...that's made a huge difference. Another slow trend over time is that on average people are working slightly less. It's a very gradual decline in the hours, towards more manageable numbers, especially as in most families, both parents are working and sharing to some extent or form caring responsibilities.

00:39:03:18 - 00:39:34:01

What's deteriorated, however, is the stability of jobs, the types of contracts, the number of people working in precarious conditions, and other employment conditions. And one of the big concerns also going forwards, especially with the rollout of new technologies, is, devices that control workers, which will have a significant or have the potential to have a significant impact on both your mental, mental or physical health.

So if you think about workers wearing wearable devices as they run around a warehouse, not having enough time for breaks, not being close enough to facilities, not having enough time to eat, that has both physical and mental consequences, having people who are being tracked on their computers in terms of eye movements know how productive they are, number of clicks per hours.

Those are all, devices that we haven't really thought about enough yet in terms of how we should regulate them.

So the role of the government is not just to mediate between the different social actors, but also to think about the future and the institutions that we need going forward, especially in a context of adapting to new technologies.

Kirsten

So the one of the issues about labor markets is it has repercussions throughout your entire life, into your family, your household and beyond. It affects your wellbeing. It affects the well-being of your family, your children and so on. And so there's a social component, or a social fallout of labor markets that we don't often take into account.

And there are many people out there who say growth is the best medicine, growth is the best social policy, etc.. Yes, growth is certainly helpful, but it's not in and of itself necessarily enough. Much depends on the regulation, on the institutions that surround that.

Outro: This episode was written and produced and edited by me, Maayan Arad (and edited by Oliver Johnson). If you’d like to find out more about the research in this episode, head to the show notes. And if you enjoy iQ, please leave us a review to help other people discover the podcast.

Join us next month, when Anna Bevan asks: Why are we having fewer children?