For years it has been read and experts have shouted from the rooftops that artificial intelligence was going to change work one day. That day seems to have already arrived. In 2024 and 2025, thousands of companies in Spain have begun to use AI for tasks that were previously done by people: customer service, resume selection, document writing, data management, technical support.
And the most curious thing is that, in many cases, the change has been so unknown or even hidden, that it has barely been noticed from the outside.
Since ChatGPT hit the market at the end of 2022, job offers for recent graduates have fallen by 32%, according to data from the Adzuna portal. While it is true that this focuses on the United Kingdom, there is no doubt that it is a true reflection of what may be happening in other similar countries.
There is no doubt that the impact is so strong that junior positions now represent only 25% of vacancies, compared to 29% before the boom from AI.
According to a Goldman Sachs report, it estimates that up to 300 million jobs worldwide could be affected by AI-driven automation, especially in economies such as Europe and the United States.
To all this we must add the fact that More than 30% of office tasks can now be partially automated with AI. And this does not refer to just factories or the use of robots in industries. This goes further and is already noticeable in all types of sectors, in all types of jobs that many believed were safe.
Plus, you probably have close cases. The friend who worked in customer service and now sees how chatbots make up a large part of his daily life. The familiar that started in human resources and today competes with systems that filter hundreds of resumes in seconds. Or the recent computer science graduate who sees that many tasks that he previously learned in his first job are now done by an AI in minutes.
With this scenario on the table, the big question is no longer whether AI is going to change employment, but which jobs will really arrive in 2026 with a future… and which ones no one dares to say out loud are in danger.
The jobs that AI cannot copy (at least for now)
To get a more in-depth idea of what could happen this year, we asked Sylvain Grande, Chief Product Officer at PayFit, who has been watching closely for years how technology is changing the world of work from within.
««Jobs that depend on emotional capabilities beyond technology are proving to be more resilient.»he explains. We are talking about jobs where it is not enough to know a lot, but where how you treat people matters.
Here we could talk about professions such as healthcare workers, teachers, therapists, social workers, caregivers or team managers. People who, in addition to basic knowledge in their field, need empathy, tact, intuition and the ability to read humans. That, today, at least, no AI really does well.
There is also another group that resists better than many think and with which Geoffrey Hinton, considered one of the fathers of AI, agrees: physical jobs. «Professions such as plumbers, electricians, bricklayers or construction personnel cannot be replaced by artificial intelligence,» comments the interviewed expert.
And then there are the positions of responsibility. ««High-level positions can rely on AI, but they do not base their main functions on these tools.»he adds. Managers, product managers, team leaders or project managers continue to make decisions that require further knowledge, context and real responsibility. AI helps, but it cannot have complete control in these cases.
«Jobs that combine deep reasoning, intuition, communication and adaptability will be more resilient»
However, everything has a B side, a more sour and complex side. Since, while some jobs hold up, others are changing so quickly that it is difficult to catch up. And some sectors that seemed untouchable are beginning to fall.
One of the clearest cases is that of basic computing. ««The productivity levels in these sectors are not the same as they were years ago.»explains Grande. AI has reduced the time needed for many tasks by 10% to 30% in junior profiles. That means that a company today needs fewer people to do the same thing as before.
Of course, this does not mean that computer scientists are going to disappear. But yes Entry work, the most routine, is the most dangerous. Programming simple functions, documenting code, doing basic testing… increasingly it is done by a machine. And this is noticeable in the market, where there are fewer and fewer junior positions and more demands from day one.
But where automation really advances quietly is in the office. The expert calls it «silent automation». Specifically, it talks about basic customer service, technical support, report writing, personnel selection… among others.
The work itself still exists, but there are fewer and fewer people behind it. And many times no one notices because the change occurs little by little, when someone retires or leaves and is not replaced.
Here’s the key point: ««This automation does not replace jobs, it replaces tasks.»explains Grande. And those who only know how to do mechanical tasks are the first to be left behind. On the other hand, those who learn to turn their work 180 degrees with the help of AI acquire power and value never seen before.
With all this on the table, and not only talking about 2026, but in the long term, the idea is to go from doing to supervising, as the interviewee comments. That is, it is no longer about producing everything yourself, but about knowing how to cooperate with machines. To think more and type less.
In the end, the big difference will not be made by artificial intelligence. It will be marked by each person’s ability to learn, move and reinvent themselves. And that, luckily, is still in our hands.