One person who knows how AI has already changed working life is award-winning Financial Times journalist Sarah O´Connor, who has just published the book ”We are not machines. The fight for the future of work”.
Read this article in Swedish on Arbeidsliv i Norden
The title it taken from the 1969-70 Swedish miners’ strike, when some of the protest was about the extensive control of miners. One of the slogans was “We are not machines”.
Sarah O´Connor has been working for the Financial Times since 2007 and is focusing on how AI and digital technology impact the power balance in working life and what it is like being a worker in a changing world.
She recently visited Stockholm to talk about her book at a seminar organised by the Arena Idé think tank.
In her introduction, she explains why she feels the need to look beyond the visions for AI that have been presented by a few men with vested interests in the industry, whose predictions are either utopian or dystopian.
The development of AI is often discussed in broad terms with focus on the distant future. Sarah O’Connor sees this as a distraction from the here and now. Through her book, she instead wanted to examine how AI has already affected jobs and industries.
“He lost everything he loved about his job”
She concludes that AI is already a reality in everyday work for the majority of employees and that the new technology has meant a dramatic change – often for the worse – in certain occupations.
One of the people she writes about in her book is a Czech freelance translator who after studying and working hard found his dream job. He started out with simple translation jobs, but after a while moved on to subtitling films and TV programmes.
This work required a deep understanding of cultural nuances and an ability to emphasise with the characters in order to convey what they would say in the local language.
As AI gained more ground, the number of translation jobs dropped dramatically. AI did the work – nearly.
All that remained for the experienced Czech translator was to use his human expertise and cultural understanding to refine AI-generated translations, for just 30 per cent of his original pay.
“He lost everything he loved about his job,” concludes Sarah O´Connor. AI could have become a tool for translators, but nobody listened to them. Instead, they have had to pay the price through lower wages and more stressful and less interesting work.
So the question she asks at the end of her talk is who controls how AI is being used, and whose voices are heard when shaping its development? How can AI be controlled by employees and not just become something you have to accept?
The social partners play an important role
German Bender is head of research at the Arena Idé think tank and a labour market researcher at the Stockholm School of Economics. He is among those who have been following AI developments and the impact on the labour market and the nature of work.
He, too, wants to bring the debate about AI back to the present and focus on how jobs are already being affected.
“So far, the debate around AI has been focused on how jobs will be impacted in the future, on AI taking over many jobs while no new ones are created. I think that scenario is very much further into the future, if it will happen at all.
“But while people speculate about the future, AI is changing jobs right now. Understanding how and why this is happening is important, not least for the social partners.
“How will existing jobs be affected and who has control over that? This is what we need to focus on, also in research” says German Bender.
Future scenarios part of the marketing
The future scenarios are usually painted by a few leaders in the AI industry, who are nearly all from the USA. They are driven by their ambition and by creating a perception that AI is where the future and the money lies.
In broad brushstrokes, they paint a picture of a future in which AI either takes over all the boring jobs or develops a life of its own and threatens humanity itself.
These are powerful visions of a world changed forever, which understandably provoke anxiety and, in turn, attract attention. What is happening here and now generates far less interest.
“These scenarios are being pushed by the very people developing the technology. And because they have access to technology that has yet to reach the market, they can see developments six to twelve months before the general public.
“Perhaps they have good reason to believe in the visions they present, but it is also about marketing and attracting more investors and driving up share prices by promising that their particular technology is the most advanced.”
“If the few people who are today’s tech leaders are given carte blanche to shape and own and industry which becomes what they say it will, that would be an incredibly worrying development.
“Not least since they are now close allies of Donald Trump and anti-democratic movements, including in foreign countries,” says German Bender.
AI already very much part of everyday life
AI is being used in an increasing number of systems, including those relating to defence and security. Some decisions are now made by AI through what is known as automated decision-making. And things are developing fast in the labour market too.
According to the Swedish Internet Foundation, 71 per cent of white-collar workers and 33 per cent of blue-collar workers in Sweden already use AI in their daily work.
Looking at overall AI use, it increased by 20 per cent between 2025 and 2026, partly because Google now provides AI-generated overviews in ordinary search results.
Has used AI at work during 2025–2026
There is no doubt AI is affecting both workplaces and society as a whole, but perhaps not in the way that was forecast. Just five years ago, some predicted large swaths of jobs would have disappeared by now.
German Bender challenged this prediction in an article in the spring of 2023, where he provided several explanations for why it will take time before technology replaces humans in the labour market.
Today, he can conclude that he was right. AI has not swept away vast numbers of jobs. Many jobs have changed, however. Sometimes for the better, sometimes for the worse.
“Good enough and cheaper” – a challenge for many sectors
How AI impacts work varies a lot between sectors. The job of a bus driver or waiter is much the same as it was ten, fifteen years ago. Yet programmers, for instance, are experiencing rapid change.
The same goes for communications professionals, translators, graphic designers or people working in the transport sector. Their jobs are undergoing major changes as a result of AI.
“In certain professions, AI might mean tasks disappear. This can be a positive thing, for instance when a doctor can leave some of their paperwork to AI and free up time for patients.
The professions most at risk are those that can be replaced by technology. AI might not do an equally good job, but employers or customers might think it is good enough and, not least, much cheaper – perhaps just a fraction of the price, according to German Bender.
Whether AI makes their work more efficient
And if these professions also have weak union representation or are largely made up of freelancers, they risk being pushed out altogether. Or they will be asked to improve on AI-generated work to make it appear more human-made, like the Czech translator, and often at considerably lower fees.
Professions and sectors with strong unions stand a considerably better chance at influencing how AI is implemented in the workplace.
One example is how the Boliden mining company introduced AI. There, trade unions and employers sat down and discussed AI and how it could be used in a positive way for the company.
What aspects could be negotiated in the same way as any other organisational change, and what could companies demand of AI providers to ensure their systems were tailored to their particular workplaces?
“Automation technology has worked quite well. Workers have been involved in negotiations, leading to improved security which in turn has resulted in increased productivity, which benefits everyone.
“And this has been achieved without increasing the surveillance of workers,” says German Bender.
Many sectors in Sweden have strong trade unions and many are usually happy to embrace new technology.
But that does not mean accepting whatever technology is introduced without question or accepting the increased risk of workplace surveillance that AI makes possible.
So it is important to learn from what is happening now and understand what different sectors need in terms of organisational change and skills development.
It is also important to recognise there are signs that young people are finding it increasingly difficult to enter the labour market in certain sectors.
“I think it is important not to buy into the narrative promoted by AI companies, as they have an obvious vested interest. Instead, employers and trade unions should always ask three fundamental questions before introducing new technology:
“‘What problem is this technology intended to solve?’, ‘Whose problem is it?’ and ‘Could it be solved in another way?’” says German Bender.
