AI-generated summary
Artificial intelligence (AI) is undergoing a significant shift known as Embodied AI, where it moves beyond generating text and images to actively performing physical tasks in the real world. This evolution was the focus of the Future Trends Forum in Madrid, which gathered international experts to explore how Embodied AI will impact sectors like health, mobility, industry, and employment. David Dorn, a labor market expert, emphasized the profound changes automation brings, especially as machines begin to replace tasks traditionally done by humans. He highlighted that automation tends to target well-defined, repetitive tasks, affecting middle-range manual and cognitive jobs such as factory work and administrative functions. However, the future also poses challenges for both low-income roles requiring complex physical and social skills, and high-income jobs dependent on creativity and problem-solving, as AI tools increasingly encroach on these areas.
Dorn addressed fears about job losses, noting that historically, automation has not caused mass unemployment but instead reshaped labor markets by creating new roles, enhancing human productivity, and generating wealth that fuels demand for diverse services. Importantly, he stressed that AI tends to automate specific tasks within jobs rather than eliminating entire occupations. This leads to a transformation of job roles, often involving upskilling and shifting workers toward more value-added activities. Thus, the rise of Embodied AI signals not the end of work, but a profound redefinition of how work is performed across all sectors.
Economist David Dorn discusses how physical artificial intelligence is changing employment at the Future Trends Forum. Which jobs are at risk? What tasks will be automated? Will there be more strike?
Artificial intelligence is crossing a critical frontier: it no longer just generates text or images, it now acts in the physical world. This transformation, known as Embodied AI, was the focus of the latest Future Trends Forum, the think tank of the Bankinter Innovation Foundation that brought together 40 top-level international experts in Madrid.
During two intense days, they discussed how this new wave of AI will affect key sectors such as health, mobility, industry and employment. David Dorn, Professor of Globalization and Labor Markets at the University of Zurich and a leading global voice on the economic effects of automation, focused on a key issue: how the world of work will change when machines not only think, but also act.
In this article, we summarize his vision for the future of employment in a world where AI not only complements, but also replaces human tasks.
Don’t miss David Dorn’s presentation:
David Dorn: “Technological development and social impact” #EmbodiedAIForum
The future of work in the face of the wave of physical artificial intelligence
David Dorn opens his speech at the Future Trends Forum with a blunt statement: “the vast majority of households with working-age adults derive their income mainly from work”. Therefore, every time a new technological wave erupts, an inevitable concern settles among millions of people:
Does this threaten my employment? Will I lose my main source of income?
Dorn frames this concern within a historical context. The question of the impact of technology on employment is not new. Since the Industrial Revolution, automation has transformed the labor market. However, the scale and speed of current changes, driven by advances in artificial intelligence—especially
Professions in the Spotlight of Automation
According to David Dorn, automation has not been random: it has followed a clear pattern, concentrating on well-structured tasks that can be accurately replicated by machines. Specifically, it has affected two large groups of occupations located in the middle range of the labour market, both in manual and cognitive functions.
On the one hand, jobs that involve physical manipulation of objects according to well-defined processes. This is the case of factory production lines, where robots have been taking on repetitive tasks for years. Dorn highlights that this automation has been made possible by the predictability of the industrial environment, noting that this trend is accelerating as robots become more versatile and adaptable.
On the other hand, there are occupations that handle information under fixed rules, such as accounting or administrative management. In these areas, artificial intelligence systems are beginning to replace routine tasks. Dorn cites as an example his university’s payment system, which already uses an AI model to automatically classify expenses according to their nature. It is not a complete automation of the job, but it is about specific tasks that were previously manual and can now be solved with algorithms.
In both cases, the common denominator is the predictable structure of the tasks: the clearer and more repeatable a function is, the more susceptible it is to be automated.
What about high- and low-income jobs?
One of the big unknowns, according to David Dorn, is whether artificial intelligence – including Embodied AI – will be able to extend its reach to occupations that until now were considered protected: both at the bottom and at the top of the labour market.
Low-Income Jobs
At the lower end of the market, there are many occupations that have traditionally been considered poorly automatable, such as hairdressers, waiters, cleaning staff or home assistants. These jobs require a complex combination of skills: visual recognition, spatial orientation, verbal communication, and fine motor skills to interact with changing physical environments. Dorn explains that, until now, this complexity has made it difficult to automate.
However, during the Future Trends Forum, incipient cases of automation in these sectors were presented, such as cleaning robots or systems for home assistance. Still, the technical challenges remain enormous. Dorn highlights the observations of expert Sonia Chernova, who underlined how every home is different: from the arrangement of furniture to the variety of objects and obstacles. These unpredictable environments mean that navigating and executing tasks by a machine remains a pending challenge.
High-Income Jobs
At the other extreme, higher-earning jobs have seemed safe because they rely on skills such as creativity, problem-solving, adaptability, and leadership. However, Dorn warns that signs of automation are beginning to be seen here as well, although in this case not so much through physical systems, but through text- or code-based tools.
He talks about the possibility of automating parts of the work of software engineers or designers, thanks to artificial intelligence models capable of generating code, images or functional designs. Although these systems still require human supervision, they represent a first step towards transforming professions that were previously considered safe from technological advancement.
Does more automation mean more unemployment?
It’s one of the big questions that hovers over every time we talk about technological advances: if machines do more tasks, won’t there be less work for people?
David Dorn tackles it directly: history shows that automation has not led to mass unemployment. From the Industrial Revolution to today, technology has replaced many tasks, but it has also created new functions and completely reshaped labor markets.
Dorn explains that there are three main reasons why work doesn’t go away, even in contexts of heavy automation:
1. New jobs linked to new technologies appear
It is the most intuitive explanation. When a new technology emerges, so do occupations directly related to its development and deployment. Dorn cites the intervention of the expert Thomas Hurd, who later in the forum would talk about the rise of talent linked to artificial intelligence: developers, data engineers, specialists in generative AI… all of them profiles that did not exist just a few years ago.
2. Technology complements human labor
Not all innovations replace tasks: many amplify them. In these cases, workers become more productive. Dorn notes that this effect can have different impacts depending on the type of occupation.
For example, if a company has two accountants and an AI system allows each to do twice as much work, they are likely to keep only one. But if researchers become more productive, the opposite can happen: by making more discoveries and developing better products, hiring more staff becomes a profitable investment.
This phenomenon has benefited high-skilled jobs above all, where technology acts as an enhancer and not as a substitute.
3. Increased productivity generates more wealth… and more demand
The third reason, less obvious but very powerful, is of a macroeconomic nature. If productivity grows, society can produce more goods and services with the same resources. This raises the overall standard of living and frees up spending capacity to meet new needs.
Dorn gives clear examples: jobs such as dog sitters, theme park employees or sommeliers do not exist because of vital needs, but because there is enough prosperity to devote resources to those services. These are jobs that arise when the economy grows and people can consume beyond the basics.
Together, these three dynamics explain why, despite the changes brought by each technological wave, employment continues to reinvent itself instead of disappearing.
Automation of tasks, not professions
One of the most important keys that David Dorn wanted to convey at the Future Trends Forum is that automation does not usually eliminate entire occupations, but transforms specific tasks within each job. This nuance radically changes the way we should understand the impact of artificial intelligence.
According to Dorn, many times the mistake is made of thinking that a job will be completely replaced. In reality, what usually happens is that certain functions within that position are automated, while others are maintained – or even gain weight.
To illustrate this, he gives a concrete example: the introduction of the ATM in the banking sector. This innovation freed employees from spending hours handing out bills. However, far from making the jobs in the branches disappear, what happened was a redefinition of the tasks. Staff began to take on more value-added functions, such as customer service or personalized financial management. In many cases, there was a revaluation of the position, promoting a process of “upskilling“ (improvement of skills) within the sector.
This task-based approach allows us to better understand why the impact of automation is not homogeneous. Two people with the same professional title can be affected very differently, depending on what specific tasks they perform. It also implies that the future of employment is not so much at stake in the disappearance of complete jobs, but in the internal transformation of roles.
In short, the advancement of artificial intelligence, including Embodied AI, does not eliminate jobs as a block, but progressively modifies what it means to practice each profession.
Profesor de Economía en Centros de Estudios Monetarios y Financieros
