# AI Workforce Strategy: Control, Trust, and Human Value

**Podcast:** FT Tech Tonic
**Published:** 2026-09-30

## Transcript

Hello, I'm Murad Ahmed, the FT's technology news editor.
Tectonic will be back soon with new episodes.
But in the meantime, here's one of my highlights from the recent FT Weekend Festival, a conversation between me and my FT colleague, Sarah O'Connor, about her new book.
It's called We Are Not Machines, and it covers a really big topic, the impact that AI is having on jobs.
We had a really interesting chat, so here it is.
Hello, my name is Murad Ahmed.
I'm the technology news editor for the Financial Times.
And I'm joined by Sarah O'Connor.
Sarah's going to absolutely hate this bit, but I have to introduce her properly.
She's an FT columnist, reporter and associate editor.
She has won every journalism prize worth winning, including the Orwell Prize and the Wincott Award.
Quite simply, she's one of the best journalists of her generation.
She's also the author.
of this, her debut book, We Are Not Machines.
It's the best book I've read on how AI is changing the world of work.
So early in Sarah's book, you write that we hear a lot about what the tech titans of the world, Elon Musk and Sam Altman, say is going to happen to jobs and the job market.
Your book focuses on what is actually already happening.
You do actual shoe leather reporting.
You talk about getting your boots dirty.
And that's the focus of this conversation here.
We're going to talk, if you want predictions of the future and an oracle, that's not what we're going to do.
We're going to talk about what's actually happening.
And it might just give you some insights as to what's going to happen next.
And maybe even be a bit of a call to arms about what you can do to affect the future.
So Sarah.
You split your book into three, mind, body and soul.
And so let's start with mind.
You start with the translation industry and you start with a Czech translator, Peter.
Do you want to explain why you start there, who Peter is and what is going on in that particular industry?
Yeah, I mean, as you say, Murad, and thank you for that introduction, the reason I wrote this book...
really is that I'm like a shoe leather reporter at heart.
What I like to do is to get out of the office, to get away from the kind of the C-suite interviews and just go and talk to people about what's really happening in the world and what that can tell us about how things are changing.
And I felt that that was really sort of lacking in this whole debate about AI.
It was so dominated, I felt, by...
these big picture predictions, you know, some of them very utopian, we're all going to end up working three hours a day and lying in our hammocks and reading poetry to each other.
Or the kind of the really dystopian ones, you know, Bill Gates is the latest, that all jobs are going to be wiped out and we're going to have to, you know, think of some sort of completely new way of organising society.
And it was kind of frustrating to me for a couple of reasons.
One is that, you know, these...
Voices are all really coming from the top.
They're coming from people often who actually have something to sell us.
Not that I think they're doing a brilliant job always on their PR, but they...
They have a kind of vested interest, and yet they're really dominating how we're trying to think about the economic impacts of this technology.
And also because, as you say, I know because I write about the world of work that this technology has already entered all of our lives.
It's entered all of our workplaces.
It has already started to transform things.
And while we were spending all of our time talking about predictions over how many jobs will be lost, all of the kind of obsession about the quantity of jobs.
I was already starting to hear that actually the quality of jobs is starting to change.
The nature of work, the way we feel about our work, whether it gives us satisfaction and if so, how.
All of that was starting to change, I thought, in quite interesting ways.
And so I really wanted to write a book that, as you say, writes about what's actually happening and gives voice to people who maybe don't always get heard in this debate.
I started with translators because they are very clearly on the front lines of this transition.
They felt it first.
And my sort of assumption had been, well, surely a lot of translators have simply lost their jobs now because we all know that artificial intelligence and previous iterations of machine translation are pretty good.
But that's actually not really what's happened.
I mean, there has been a kind of...
decreasing demand for translators over the years but a lot of translators do things that fundamentally the machines aren't quite good enough yet to do and so that leaves the question of well how are their jobs changing and is it for the better or is it for the worse so peto who you talked about so he translates subtitles He lives in the Czech Republic.
He translates the subtitles of like Netflix shows, Disney shows.
He absolutely loves this job.
He says it's his dream job because you get to watch a lot of TV.
He loves TV.
But also it's like it's a very creative job in ways that I hadn't really thought about.
So yeah, Petra was telling me all kinds of things, kinds of decisions that he has to make, judgment calls.
So one would be...
In Czech, there are two different ways of saying you.
There's the formal you and the informal you.
It's true of lots of European languages.
But obviously, in English, we just say you.
So if you're translating like an American show, you're the one who has to decide in any given scene, well, what...
tense would they be speaking in here?
How familiar are they with each other?
Are they friends?
Are they colleagues?
Sometimes that might change at some point in the course of an episode or a season.
And that's a kind of human judgment that you have to make.
You know, the other thing is jokes.
So like jokes are incredibly difficult to translate, as you might imagine.
A lot of jokes are like play on sort of shared cultural knowledge in a particular country or their wordplay, you know.
And so...
They're really hard to do, but they're really enjoyable when you kind of nail it.
And that's what he said to me.
There's this profound sense of satisfaction when you figure out some way of translating a joke that plays well, like on the east coast of America, in a way that's still funny in the Czech Republic.
My favourite example, actually, of the genius of translators is the Asterix books.
So obviously they were written in French, but some of the English translators are just amazing.
So yeah, one of my favorite examples is Obelix's dog, who in the French he's called Ede Fixe, which means obsession.
But the English translators called him Dogmatics, which if anything is an even better joke, right?
So it just goes to show how creative this job is.
So what's happened to Petr?
How has AI changed the job for people like him?
What's happened is that a lot of people who do that are freelancers and agencies will kind of intermediate this work between the studios and the freelance translators.
The agencies have brought in something new called machine translation post-editing.
And what that means is that the agencies will take the subtitles in English.
they will get a machine to translate them, and then they will give that machine copy to the professional translator, who's now told to tidy it up, to check it for accuracy, maybe make it sound a bit more human.
But they're asked to do this at twice the pace and for half the price.
And so what happens is a job that...
used to be quite fulfilling and creative, has become more intense.
You know, he feels like he has to rush through this work now because he has to still make his kind of his weekly or his monthly income, but you have to rattle through it faster.
But not only that, he said that it actually loses the creative part of it.
And he said that, you know, when someone sort of presents a solution to you and you're looking at that, it's very hard then to kind of think of your own solution.
Like it's already sort of set the...
the frame of reference for you.
Plus you're in a hurry, right?
And so basically what he said was a job that used to be really profoundly satisfying and creative had become more repetitive, more mechanical.
You know, he started out working a production line in like a window and doors factory.
He was a manual laborer to begin with.
And he said it felt a bit more like that kind of job.
And the reason I started my book with that example, although it's a bit depressing, is that I think it's really important because it...
This is not at all what we were promised this technology would do, right?
We were told that this technology would allow us to become more human, that we could offload the repetitive stuff, the mechanical stuff, the stuff that's dull or dirty or dangerous, and that we would be freed up to concentrate more on the parts of our jobs that rely on our human skills, like creativity, like interpersonal skills, all of that sort of thing.
And actually what...
The translators were telling me was that what was happening to them was the precise opposite of that, that they felt actually more like machines themselves and that they were kind of being crunched into a system where they had to kind of plug the machine's adequacies, but at a machine-led pace.
But I will tell you, not all the stories in the book are as depressing as that.
Because there are some good things happening too.
We will get to the optimistic things, but let's set the tone and depress people first before we bring them up again.
But what you find in the translation industry is that they also get paid less as well.
AI has driven down.
the value of their work, at least in the market as well.
And there's a question that one of the translators asks is, what do you do when you see your profession start to fall apart?
Now, AI is affecting different industries at different rates and different ways.
But if you're in one, like the translation industry, where it's very clearly eliminating either roles or making them less satisfying, as you're describing, what do you do?
How do you respond to that?
Yeah, I think that's a really good question.
And it actually kind of allows me to mention something that really struck me when I was doing the reporting for this book, which is that, you know, a lot of those narratives that I talked about at the start, both from the kind of tech chief executives, but also from a lot of economists and probably a lot of journalists like me, and I'm sure I've done this myself in the past, is that...
We've often sort of thought about this technology as something that's just going to happen.
Like, you know, this wave is coming at us.
The tsunami of change is coming at us.
And what's going to happen is there will be winners and losers.
You know, we hear this a lot.
And then we're trying to map out, like, who will the winners be and who will the losers be?
Who's going to be most exposed?
Who's most affected?
But actually, when change comes to you, when it enters your door, when it enters your workplace and it starts to threaten your job or the quality of your job or how you feel about your work, people don't just like lie back and go, oh, well, I guess, you know, that's the end of me.
I'll just wait for someone to give me a universal basic income in 20 years time.
Like, that's not what people do.
Like, all of us respond.
You know, we all do things.
And I think that that is a big part of the picture that...
economists are missing is that it's not just what the technology can and can't do.
It's also how people will and won't respond to it.
So yeah, none of the translators were just like lying back and taking it.
So Petra, for example, he had started to train in different areas of translation that were less vulnerable to...
AI disruption, so doing something called localization for video games, which involves quite a lot of kind of cultural knowledge as well.
He was also working on doing subtitles for the deaf and hard of hearing, which again felt as if it would be less vulnerable.
But he did say that he felt as if everywhere he went that AI was kind of following close on his heels and he was sort of trying to get out of harm's way.
Yeah, I mean, the...
Because I speak to a lot of tech industry people, and the phrase that I've started to realise is quite Orwellian is human in the loop.
They'll say you have to keep a human in the loop.
Whenever you hear that, that's a massive red flag.
That's trying to turn us into an army of fact checkers while the machine goes away and does it.
Always question whenever you hear that sort of thing.
We're more than in the loop, we are the loop.
So in the interest of getting your boots dirty, you also went down some Swedish mines.
Explain why.
Yeah, so as well as writing about the white-collar world, I also wanted to write about blue-collar work and how automation is changing that.
Because again, I feel like that's something that has had a lot less attention, but it's also going to bring in some quite profound changes.
Not least in terms of self-driving cars and trucks.
We're starting to see self-driving cars on the road.
Self-driving trucks are being trialled.
but they're not really yet sort of commercially ready because there are an awful lot of kind of very...
sensible regulatory caution around the idea of self-driving trucks.
But there are places where self-driving vehicles have taken off massively.
And that is in places that most of us never see.
Like, for example, mines, other kind of big industrial sites, which aren't, you know, they're not driving on public roads, but they're actually very useful in places like that.
So I wanted to go to Sweden for a couple of reasons.
One is that this mine that I went to, it's right in the far north of Sweden, near the Arctic Circle.
It's very technologically advanced.
They're very pro.
new technology in this mine and they brought in self-driving trucks and loaders underground which means that the miners now are not driving those trucks themselves in these like pitch black passages that I went down into they're now sitting in a control room and they have sort of panels of screens And they're watching what the trucks are doing, and they can intervene if they have to.
They've got a joystick, they've got pedals.
But for the most part, their job is quite different.
They kind of look a bit like investment bankers with all of their big screens, except they're wearing overalls instead of gilets.
I was going to say suits, but no one wears suits anymore.
It's all about the gilet these days.
And I want to know, how has this gone for the miners?
How do they feel about this?
Is this an improvement?
One of the hopes, as I mentioned, of...
this new technology, is that it will allow us to offload some jobs that actually maybe we don't want to do anymore.
And mining has long been like a profoundly dangerous, horrible job.
It doesn't mean that people haven't valued it because it has provided some good livelihoods over the years.
So it presented me with an interesting question.
Second reason I was interested is because Sweden has quite an interesting approach to industrial relations that's very different.
to how it works here or in the US.
And I feel like the debate about AI and how it will impact the world of work has really been dominated by America and the American labor market.
We're all looking at the US labor market, which makes sense.
That's where the big tech companies are.
There's 300 million people in the US.
There's a massive economy.
It matters.
But a lot of other economies deal with technological change in very, very different ways.
And so that means that it might affect the outcomes.
So I wanted to know how it was going in somewhere like Sweden.
And the answer was quite well.
So the miners, they're quite happy sitting in front of their screens.
You know, one of them said to me, yeah, like...
It's less noisy, it's less dangerous.
I'm listening to Spotify now because we've got Wi-Fi in the mine.
I'm listening to the ice hockey on the radio.
I've got my mobile phone with me so I can be in touch with my wife and my kids.
Before that, you would just be cut off from the world for your eight-hour shift when you were underground.
So I think this has been an improvement in the quality of that job.
But that doesn't mean that there weren't lots of quite interesting and sticky points along the road where it could have, I think, gone badly.
I'll give you a couple of examples.
One was, before you even get going with the self-driving trucks, you need to get Wi-Fi working in the mine.
And once you've got that working, it enabled the company to think, well, now we can have a real-time positioning system.
So we can now know where all the miners are at all times, which from a safety point of view is a great idea.
And a lot of the miners saw that, particularly, so this company owns lots of mines.
And those miners that worked in mines where there had been fires.
in the recent past were really pro this technology but some of the other ones were really worried about it because they were like well could this not also surreptitiously become a surveillance mechanism could you not then start tracking where we are and what we're doing and how long we're taking?
One of the miners literally said to me, we didn't want to be treated like Amazon warehouse workers.
We didn't want people timing how long we spent in the bathroom.
But because it's Sweden, they have this thing called co-determination.
So whenever a new technology comes into a workplace and management wants to implement it, first of all, they have to sit down with the unions and they have to talk about it.
So management still has the ultimate prerogative.
It's still for management to decide what to do, but they have...
to at least negotiate and talk it through.
And the unions do have quite a lot of power to hold up the process if they want to.
And I kind of thought, coming from the UK, where often the relationship between unions and management is kind of antagonistic, I thought this would be just a recipe for total sclerosis.
Like, how would anything get done if you have to sit and negotiate every new bit of technology with the workers?
But actually, in Sweden, it had gone quite well.
And I think that's partly because they've been doing it this way for...
decades, right?
So they've got muscle memory of how to negotiate about things.
So I sat in a room with the head of the automation program, the head of the white collar union and the head of the blue collar union.
And it was clear that these three men, although they were technically on sort of opposing sides of the like capital labor divide, they know each other, you know, they've worked together for a long time, they respect each other, and they know how to negotiate with each other.
So what they'd done with this real time positioning system was they'd come to a compromise.
And the compromise was this.
They went back to the technology company that had designed it, and they said, can we pay you a bit to modify this system, to add a new feature so that we can know where everyone is at all times, but not who they are.
So everyone will have an anonymized number tag.
And there will be certain situations where we'll de-anonymise it.
So both the management and the unions will have a password.
And when they put it in together, we can de-anonymise that data if there's an emergency or if there's an accident that needs to be investigated.
But for the most part, it will be anonymised.
And this was a really good solution, it seemed to me, because it allowed management and workers to reap the benefit of this technology, which was safety.
But without...
eroding that level of trust that already existed in the workplace.
And I don't think that that would have happened if they hadn't talked about it first.
For me, this is an example of how actually, although you might think that negotiating over technological change could slow it down, in some ways it kind of just allowed it to move forward faster.
Another example was that the management wanted to put the control room on the surface.
They were like, hey, you don't have to go down into the mines anymore.
That will be more efficient.
It will be safer.
But the miners really wanted to stay underground for a good reason, which is you get paid more if you work underground.
And no one's going to go along with something that means a pay cut.
So again, they came to a compromise.
In fact, the head of the Blue Collar Union described it to me as a very Swedish compromise not to increase the conflict level, which was that they put the control room halfway down the mine.
So I went to visit it.
So you go down, there's a canteen there.
So you walk through the canteen where the miners have like heated up their curries in the microwave.
You go through that and then that's where the control rooms are, where the miners are sitting and supervising.
And again, like from the outside, like...
As a journalist or as a management consultant, you might say, this is so inefficient, this makes no sense.
But I would argue that, in fact, it's more efficient because it allowed them to just get on with it.
It allowed them to implement that technology to reap the benefits of much higher productivity and safety without getting kind of caught up in like a long-running battle with workers over something where they felt that they were going to lose out.
We're sometimes invited to think of technology as something that you must either resist...
or acquiesce to.
But I think this Swedish mind made me think that there are actually, there are other paths forward as well.
There aren't necessarily either of those.
And the chapter on this was really enlightening because you were talking about a completely different cultural way of thinking about management.
There's a quote that you...
start the chapter with from Elon Musk.
And he says, I disagree with the idea of unions, perhaps for a reason that is different than people may expect, which is, I just don't like anything which creates kind of a lords and peasants sort of thing.
And this is the argument that's made.
You know, actually, you're empowered by not being part of the union.
You're not getting top-down instruction.
But I think what I came away from, and I've told you this before, Sarah, I don't see you as a polemical writer you report on things dispassionately but as much as I detect any kind of like an argument that you're trying to steer us towards is that you believe that there should be this rebalancing between the worker and the manager and it's what we all need to move towards is the sense of there can be a world where we talk to each other with trust rather than the sense that what matters is control.
Yeah, I think that's right.
And I don't think it has to always be in the form of like employers and unions.
Basically, I found some really depressing stories like the ones of the translators and some quite heartening ones.
And I could have kind of concluded by saying, well, there's going to be winners and losers.
You know, there's opportunities and dangers.
But I think that's really a cop out because it's true, but it's not helpful.
You know, it doesn't ask the harder question, the helpful question, which is what makes the difference?
You know, why is it going well?
in some places and for some people and badly in other places for other people.
And so I tried to kind of challenge myself to try and answer that question as best I could.
And what I came away with was that what was making the difference was who had a seat at the table when these technologies were being implemented in the workplace or even designed?
And do you as the worker have...
control over when to use these tools, which tools to use, and how to apply them to which parts of your jobs.
Because we all know our own jobs.
We know which parts could do with a bit of automation or a bit of fast-forwarding through.
We also know which parts of our jobs we profoundly value and don't want to offload.
But when that is imposed on you from above, that's often where it's going wrong.
you know, where people are losing things that they actually really cared about or where they feel as if they're kind of having to jam like a round peg into a square hole, you know, that you're being told to use a tool that actually isn't quite right, isn't really doing your job properly, but you're under pressure to use it anyway because management has decided that this is the future and that everyone will be left behind otherwise.
And then that leads to a lot of kind of resentment.
It also leads to lots of extra work because either you or your colleagues have to kind of check it, tidy it up.
make it better.
So I think that that's really where the difference is.
And so I don't think it's about being pro-AI or anti-AI.
I think it's about being pro-giving people control and power over it.
You end your book with this rather beautiful story between a nurse, Jennifer Bergkamp, and someone towards the end of their life, Tony Ewell.
Do you want to just talk about their relationship and why...
That tells us a lot about what we could learn here about end-of-life care because it's something that's frankly on a lot of people's minds.
Yeah, I wanted to write about care because I think it's one of those areas where a lot of us are kind of told that we need to kind of just accept that automation is going to be necessary.
You know, we all know that...
populations are ageing, that demographics are not looking good, that there are going to be fewer working age people and more people that need care.
But also that these are the sorts of jobs that nobody wants to do.
You know, there are in the UK...
There are massive shortages of people who want to work in social care.
Turnover is very high.
And so the argument is often made like, well, this is the kind of area that we need to apply technology to.
You know, we need to figure out how to use robots to help care for people when they're old.
And we need to, maybe we can use chatbots to help keep people company.
And maybe we feel slightly uncomfortable with this, but fundamentally, like there's no other choice.
This has to happen.
And so I wrote this chapter, really, I think I called the chapter Questioning the Premise.
because I think that this organisation that I went to visit, which is in the Netherlands, is an example of why sometimes there are other ways of answering that question that don't necessarily have to involve technology and automation, or at least not in a really all-encompassing way.
So this organisation that Jennifer, the nurse, works for is called Boertsorg, and it was founded by a guy who had been a community nurse in the Netherlands and had watched that whole system change over the years.
gone from something where the nurses had quite a lot of autonomy, they were based in their own communities, to something quite similar to what we have here.
So like big companies involved, private equity involved, lots of layers of management, and that the job of being a home care nurse, someone who goes and visits people in their own home to take care of them, had changed into something where it had almost been sort of turned into a bit of a sort of production line itself.
And so you might end up, as an elderly person who needs care, you might end up with sort of five or six or...
seven or eight different people coming to your house every week to perform different tasks for you.
And they might all have sort of 15 minute time slots in which to do it.
And this was seen to be more efficient because, for example, why would we have a qualified nurse come in just to give someone a shower?
You know, you don't need someone that's...
paid this amount to do that.
We could get a lower qualified person to do that thing.
So let's break it all up and we'll get different people to do the different tasks and it will be more cost efficient.
And Yost, the guy who founded this company, thought that something had been lost there and that there was a different way of doing it.
So he started this new organization where basically nurses are based in very small teams, like six to eight nurses.
They are...
fully autonomous.
So they live in one community, they care for the people that need help in that community.
They do everything themselves.
There is no management.
So they decide where their office is, they find their office, they decide their rotors, they decide everything.
They go and meet their clients and figure out how much care they need and how they can provide it.
So, you know, there might be one or two people that go and visit one client.
And so you really get to know...
that client very well, which was the case with Jennifer and Tony, who was the man that I went to.
So I hung out for a few days with Jennifer and her team to see how this was going.
And what she said to me was that while you might think it's less efficient to have someone well-qualified like me to just go and make someone a sandwich, for example, in the morning, she said, actually, it's not.
because you're not just there to make the sandwich.
You're there to see the life.
You're there to see how they're doing.
You're there to check in on them.
And, you know, she gave me the example of someone who she'd gone to make her a sandwich every day for a week.
And three days in a row, she'd found the sandwich uneaten in the bin.
And she said, well, that could mean several things.
That could mean she's lost her appetite because she's dying.
You know, the appetite is the first thing to go when you're dying.
It could mean certain kinds of stomach cancer.
Or it could mean that she hates the kind of cheese I've been using.
But she said if...
A different person had been going in every day.
None of that would have been noticed and no one would have been able to have that relationship of trust and a bit of banter where they could just talk about it.
Same with giving someone a shower.
You're not just giving them a shower.
You're seeing how they're doing.
You're noticing if there's an increased frailty or something doesn't look right.
And therefore, you can spot problems much earlier.
and address them faster.
And so what this model has basically done is lower the cost of care.
They actually provide fewer hours of care per patient, but satisfaction is much, much higher.
I was curious about how this felt for Jennifer, like for nurses, and she said that she felt it was the most beautiful job in the world.
She's so...
angry when people say that there's a shortage of people willing to care for people because she said actually if you do it right if you set it up right if you organize it in the right way it is a profoundly satisfying and beautiful job and the relationship between her and Tony I sort of witnessed that you know they weren't just like a task provider and a task recipient to each other you know they knew each other She's an artist as well.
She draws pictures, beautiful pencil portraits of the people that she cares for.
In the process of doing that, she talks to them, she gets to know them.
It was a reminder that actually, yes, we could just say, these are terrible jobs.
No one wants to do them.
The only answer is automation.
Let's get the robots in.
Or we could say, is there a different...
solution here like is there a way of making a bad job into a better job in a world that's actually going to need good jobs for humans and where one day we're all going to need someone to care for us and would we rather it's someone like Jennifer would we rather it's a robot supplemented by a chatbot yeah and there's a bit I mean sorry for the spoiler but there's a bit that I felt something like cat to the back of my throat as I was reading it because you explain how at the very end of your life When you are about to die, a lot of people hang on if a loved one is near them.
This is so deeply human and the nurses often just say to the loved one, I think if you just left the room, now will be the time.
And I don't see droids being able to work that moment out anytime soon.
Yeah, and also...
You know, while I was writing the conclusion of the book, it was like at a time when there was a flurry of tech chief executives saying, oh my gosh, you know, these new models are so amazing that soon AI will be able to do everything humans can do better.
There won't be a single thing that we will be better at than the machines.
And, you know, I thought about Jennifer and how she can sit there with someone when they're dying and know what's the moment to go.
and hold someone's hand and say, it's okay, now is the time.
I thought, A, I don't think a robot will ever be able to do that.
B, I don't think it should.
We're asking so many questions at the moment that start with the words can and will.
Can a robot take my job?
Will AI be able to do this, that, or the other?
But we're not asking that many questions that ask with should or ought.
What do we actually want?
technology to do and what do we not want it to do?
You know, and that is for all of us to decide.
We don't just have to accept that it's going to sweep into every area of life.
There are some things that we can all decide.
Actually, that is the preserve of humans.
And that's the way we want it to remain.
And thank you, Sarah.
Thank you, Murad.
That was me, Murad Ahmed, speaking to Sarah O'Connor about her book, We Are Not Machines, at the FT Weekend Festival.
My thanks to Sarah.
You can keep up with her reporting at ft.com.
and by signing up to the AI Shift newsletter that she co-writes.
Tectonic is produced by Edwin Lane.
The executive producer is Topher Forges.
Sound design by Breen Turner and Sam Giavinko.
The FT's head of audio is Flo Phillips.