It was a sample of “Ia-Pocalypse” that was going to take everything to its path. In February 2024, Klarna’s Swedish specialist in fractional payment announced his decision to eliminate 1,800 customer and marketing advisors to replace them with chatbots supplied by the artificial intelligence of overwhelmed Chatgpt.
Sebastian Siestkowski, the Fintech CEO, estimated that AI had become so powerful that it could reduce the work of 700 human agents, with a resolution rate of less than 2 minutes at request against 11 minutes with humans.
Some (including investors) saw in their statements the end of the relations of subcontracting customers, such as the formation of French telepers. The action had decreased by more than 14%.
Fifteen months later, all rear machines. The company wants to give employees in meat to advise their customers. The quality of defective advice, the impression of speaking with a rigid machine disappointed the users of the credit platform.
“We believe that this shows tacit recognition of the fact that customers always want effective and emotionally intelligent human support,” analysts of Royal Bank of Canada wrote, in response to these statements.
The Swedish company is not an isolated case. Air Canada’s company also retropped after receiving the order to reimburse a client who had been promised by a chatbot a false reduction after the death of a loved one. On the side of the DPD operator, we bite our fingers since a conversational agent insulted a client before qualifying his own company “worse delivery to the world that would not recommend anyone.”
We can still mention McDonald’s, who put (at least temporarily) the crime about the deployment of his vocal assistant that was used to order the units of his restaurants. The orders were full of errors, the wizard does not include the accents and words of Argot.
“Why so many job offers for translators?”
Numerous comments that take the opposite vision of very pessimistic forecasts on the hundreds of millions of works that had to do expiration worldwide.
The first work had to be the translator, according to the calculations of two researchers at the University of Oxford. However, employment data in the United States shows that in 2024 the number of employees in this activity sector increased by 7%.
Perhaps the effects of AI on employment really refer to graduate young people, especially in a varied number of sector?
This is what the economist wanted to verify when analyzing the unemployment rates of future white passes. In fact, the unemployment rate of young American graduates is rather in the United States. Now there is even 2.2 points above the average, all combined age groups.
Except that, the magazine points out, this increase began in 2009, long before the arrival and the IAD in companies. And above all, it is still very contained since the unemployment rate of young American graduates of higher life remains at a very low level of 6%.
The same “disappointment” for the employment rate of “white passes” as a whole. The categories of works of managers in sales, marketing, IT and more generally in offices offices have not decreased in total employment in the United States according to the figures of the Office of Labor Statistics. They even represent almost 63% of total employment in 2025, their highest level ever registered (excluding the Covvid period that had seen the unemployment of low -rating professions for a short period).
The same trend is working in all OECD countries, including those of the euro zone. Employment rates (participation in working age that have a job) have been at the top since the beginning of the statistical series, or 70.3%. And wages have also increased, 3.1% in 2025 in the euro zone according to the European Central Bank (ECB). So many signs that contradict a replacement of employees with smart machines.
“The AI will not take your job but …”
How to explain that AI, despite devastating ads, does not exert down pressure on employment in developed countries?
First, because companies use very little. Less than 10% of them have adopted this type of technology, according to GPZero.
“Often, employees do not adopt new AI technologies and do not take advantage of it, but we really do not know why,” explains Natalia Vuori, assistant professor at the Aalto Finnish University who conducted a study on the adoption of these technologies in companies.
His team conducted an investigation for more than a year in a 600 employee consulting company who had massively adopted artificial intelligence tools to assign missions to the most qualified employees. The machine had to analyze employee activities and communications to determine skills mapping. Except that employees have gradually adapted their behavior to provide information about their real skills, but in those who probably benefit them in the future. It is not relevant information about your real abilities.
Therefore, AI has become increasingly inaccurate in its results, says the study, feeding a vicious circle where users have begun to lose confidence in their abilities.
The other advanced explanation is that companies that really adopt these technologies do not rule out employees. The few examples are very publicized, but in fact these cases remain very rare. When companies are not a completely rear back machine, such as the Swedish Klarna.
AI technologies are often additional tools provided to employees to improve their execution speed and in the end its productivity. In other words, AI is today a co -piloto (as the correct name given by Microsoft to its assistant) that improves the performance of those who know how to take advantage of it.
Source: BFM TV
