I think we've done a poor job so far as leaders of articulating: how is this technology going to help the company from a strategy perspective? How is this going to benefit our workers, our employees? What specifically should they expect to see in terms of their work lives and the part that's going to be impactful for them in a positive way? So articulating that with clarity and confidence is absent in today's conversation and frankly in many organizations.
Welcome to Meet the Leader, the podcast from the World Economic Forum that speaks to the top leaders who share how they're tackling the world's biggest challenges. In this episode, I'm talking to Jonas Prising, the chair and CEO of ManpowerGroup. I'm Robin Pomeroy at the World Economic Forum, and this is Meet the Leader. Jonas Prising, how are you?
I'm excellent, thanks, Robin.
You wrote an article on Forum Stories on the World Economic Forum's website: "Over the next 5 years, the defining challenge will be redesigning work so people can grow with it." What did you mean?
Well, we talk a lot about and are frankly mesmerized by the speed of the technological evolution and the speed of that innovation. What we forget is the human aspect of how work gets done. And in our experience working across 80 countries globally with millions of workers deployed across hundreds of thousands of companies, what we know is that technological innovation really depends on human adoption. And as we think about the work and how work will need to be redesigned around AI, that's how we can see the evolution accelerating from what today is really AI being used in pretty narrow use cases as a tool versus a transformational change of end-to-end value creation processes where work will change dramatically and fundamentally as we look into the future.
So what we're seeing is AI kind of bolted onto the existing workplace. What you're saying is that existing workplace is going to change. Can you think of some examples, or is it too early to tell what that redesign will look like?
Well, it's a little bit early to tell, but I think we can see the early usage of AI is as tools on specific use cases and tasks. So if you think about a job, a job consists of maybe 18 to 25 tasks. Today companies are deploying AI to solve a specific problem within a task. And so as we see the narrative evolving around the speed of technology is moving very quickly and adoption is slow, it's because it's only being deployed in specific tasks and use cases. And over time, to see the true value of AI, we believe organizations will have to rethink how work gets done across their entire value creation. That of course is a transformation that takes time, requires significant investments both in technology infrastructure and frankly into change management, because ultimately what's going to be very important is going to be for workers and employees to understand: what benefit does this have for myself, and how is this going to benefit the activity that I enjoy doing today so that I can do more of it and less of the things that I don't enjoy doing as much?
Because a lot of those workers will be concerned that the transformation from AI in the workplace is to get rid of all the human jobs. You don't believe that's right?
No, I don't believe that's right. You listen to the technology companies, they will say, you know, job losses of X percent of the workforce is what's to be expected, so brace yourselves—frankly, a very negative narrative. From our experience, that's not how this technological innovation and revolution will play out. We will eventually find the right mix between what the technology can do to improve the performance of the organization and the performance and the value for employees as well. Based on the narrative today, it feels as if everything is going to change tomorrow. We think it's going to take time because it involves process change, change management, and it involves trust. And today we can see people aren't that optimistic about the impact of AI to their jobs and to themselves in their professional lives.
I wonder what needs to happen to change that trust, and maybe you're seeing organizations where they are doing a good job in bringing the workforce with them.
Well, I think a lot of it will depend on the leadership. So I think we've done a poor job so far as leaders of articulating: how is this technology going to help the company from a strategy perspective? How is this going to benefit our workers, our employees? What specifically should they expect to see in terms of their work lives and the part that's going to be impactful for them in a positive way, and generate that trust in the future and the confidence that proceeding along this route is ultimately going to be beneficial for myself and for the organization that I work in? So articulating that with clarity and confidence is absent in today's conversation and frankly in many organizations as well. We've gotten way ahead of our skis in terms of the innovation driven by the speed of technology, and we're forgetting that the actual adoption and the implementation depends on people's trust and confidence that this will do really good things for them in their working lives.
Your research finds that a human reviewing résumés remains the best hiring resource, because those of us who've looked for jobs, we fear it's an AI that's going to decide whether or not our resume is good enough to get us an interview. I mean, what are humans doing better at looking at a resume and deciding maybe this is a candidate for a job?
Well, some of the unintended consequences that we're seeing with the adoption of AI is not only do companies adopt AI, individuals are adopting AI as well. And frankly, the hiring experience today for many job seekers is inherently very unpleasant and unsatisfactory, because they are sending out applications to a faceless recipient, not getting a response or any feedback that is useful to them, and sending these missives out into the dark space. So the job seeker experience frankly with the technology has not gotten better; it's gotten worse.
From an organizational perspective, there's been a massive increase in applications to any particular job posting. And that's because job seekers are deploying agentic AI and AI to automate their job applications. They match those job applications and those resumes to search words that they are seeing in the job ads. So what we're seeing at Manpower is a tremendous increase of applications and a tremendous uniformity in the applications because everyone is using the same large language models to optimize their resumes on digital media to look as if they're suitable for the job.
So what that all ends up doing is that the human judgment and the conversation that we have as recruiters with the candidates that we are talking to becomes the most important hiring tool that we have. So paradoxically, the technology may increase the speed, but at this point doesn't increase the accuracy. And that human element really is the great benefit that we can provide both to the candidate and to the employers that are looking for the talent, because we have a lot of experience, we can judge this in a way that truly adds value to both stakeholders. And we've been very purposeful in designing ethical AI inside our company so that any decision—the ultimate hiring decision or recommendation—always lies with the human. And so a lot of the transactional work that we used to do, we don't do anymore. We do more of the judgment, the interviews, the interaction between the candidates to find out what the perfect fit for that individual would be.
How do you hire? Imagine we're in a job interview and I'm applying for a job for one of your clients. What kind of questions would you ask me? What would you want to hear from me to convince you to go through to the next round or to give me the job?
It's a great question, and frankly we would do this as we've done with other technological evolutions. We'd ask you, Robin: "So what do you use AI for in any aspect of your life? Are you familiar with various large language models? What are you using it for? Why do you find it useful? What do you not find it useful for? What are your experiences? Where were you really disappointed when you were leveraging this tool? Where were you really positively surprised about? And how do you think that can apply in the future as you come and explore work opportunities at Manpower?"
Crucially, you said in those questions: "What disappointed you? What is it not good at?" So if you had a candidate coming in saying, "Yeah, it's great. ChatGPT does all my work for me now. I'm going to come in and I'll be done by lunchtime," you want some critical thinking there.
Yeah, critical thinking is another element of exactly that: what is it that it does well and what does it not do well? In our own organization, we try to be very clear and be very explicit around what can go wrong and why we want human judgment and assessment at the end to be the determining factor whether a candidate moves forward from a hiring opportunity perspective, because there are many things that can go absolutely wrong if you blindly trust the technology.
Do you think—I mean, we could put a phone here now, record our job interview, and when the candidate's left the room, say, "Okay, Claude, yes or no to this one." Would you say that's a good—I mean, you could do that and choose not to. Would you do that, do you think?
Yeah, we wouldn't do that. Ultimately, whether you or I fit into an organization is so much more than the technical background and the factual background. It's: does Robin fit into the culture here in this office? Would he interact well with his boss or not? Is the culture conducive to his personality? With job seekers, for instance, it's very important to tell them that ultimately the assessment that I make of your qualifications will be by myself, aided by technology. So I've gotten some good information, but the decision will be made by myself or my colleagues. So to be very explicit and upfront, because I think a lot of job seekers and individuals at large will be disappointed if they think that we're having this conversation, but ultimately the decision's already been made by a machine. That doesn't feel very good and it's also potentially really not guaranteeing a very good outcome.
A lot of countries are seeing population decline. What do leaders of companies need to be thinking about, and leaders maybe of governments as well and other organizations, to handle what could be really fundamental change over the next 5, 10, 20 years?
Well, I think the good news is that there are people and human potential that we can unlock in new ways. This forces us to unlock that human potential, and we're not relying anymore on population growth providing us with a new flow of workforce because there are pockets of our societies where people haven't been able to participate either in globalization and presumably in aspects of this new technology innovation. There is then the pressure of trying to make sure that we bring in talent pools that have been previously overlooked or have had difficulties getting into the workforce—now making extra efforts and understanding how we can bring them in because we need them so that we can continue to grow the prosperity of a community, an organization, or a nation.
So the negative aspects of slowing population growth—which, of course, historically has always been key, population growth drives prosperity growth along with productivity improvement—the fact that we are now coming into a situation where a lot of nations are no longer having the population growth as a driver of economic growth is a big change. And I think both the promise of the technology evolution that we're living through right now and the opportunity to tap into the talent pools that haven't had the same opportunities to participate forces us to make greater effort to integrate them into this evolution so that they can also have meaningful and sustainable employment.
What kind of parts of society are you talking about?
Well, first of all, the most underutilized resource in our workforce is female participation in the workforce. High in some places, mostly much, much lower than the 50% of the population that it represents all over the world. So the female participation rate is a tremendous opportunity for us to bring in skilled, motivated, and engaged talent. Of course, we have to facilitate that participation more than what we are doing today, be it through childcare, afterschool care, flexibility so that women can participate and they still carry a lot more of the burden.
But then you also have other aspects of the population: long-term unemployed individuals who have had great difficulty getting back into the workforce. We need to redouble our efforts and bring them in. And of course, also youth unemployment, which tends to be at twice the average unemployment rate in any industrialized country and in emerging markets many, many times higher unemployment than that. And that is a key opportunity that we have to ensure that we bring in young people in early so that they can contribute and feel the benefits of employment or work.
And that's another challenge that might be exacerbated if we do AI wrong. There's research suggesting already those early careers, those early roles, those jobs are being done by AI, and there's a risk of a generation of people missing out on those early roles. So that youth unemployment could even be exacerbated if we're not handling AI right.
It could, and I've read these reports and the studies. Our own data, however, would indicate that at this point we think that increase in youth unemployment and the difficulty of finding employment for college graduates and high school graduates is more related to the economic cycle than it is to the structural change. Of course, you'll always be able to find certain pockets, say young coders with no experience having more difficulty now being hired because we can see coding is one area where AI is having an impact in demand, although it's still in most countries one of the top three or four roles that companies are looking for—so less intense demand, but still very high demand.
But aside from those pockets, we think this has much more to do with the economic cycle. Employer confidence is low. Economic uncertainty is high. Geopolitical tensions are high. That means employers pull back. Who are they hiring in times like this? Individuals—talent that have the requisite skills to become immediately productive and contribute to the areas that they want to accelerate. They are less likely today, as in the past, to hire younger people that don't have the experience and that cannot be as productive as quickly as they want the talent they bring in today in this environment. So we think this is more of a cyclical issue, an economic cycle issue, an employer confidence issue than any impact of AI. So it may have an impact over time, but as with any prediction of what's going to happen with AI in the labor markets, I think it's way early to call, and whoever makes those calls today is going to be mostly wrong.
The World Economic Forum did a Chief People Officers Outlook report that finds that geopolitical forces, something you just mentioned, are directly shaping talent strategy with knock-on effects to migration and visa restrictions, localization policies, cybersecurity. There are certain policies maybe that are making it difficult to hire across borders. Are you seeing that? What do you think of that finding?
No, I think that what the World Economic Forum's research is showing is clearly reflected in reality. So companies are much more cautious about cross-border hiring at this point in terms of knowing whether the person that is on the other side of the digital screen is actually Robin or is it somebody else, or an AI reflection of Robin trying to find employment on behalf of somebody completely different for non-nefarious or nefarious uses.
That's already happening?
That's already happening, and I have seen those attempts with my own likeness, my own voice that has been used from interviews like these to have actual conversations with my colleagues. And the only reason they can determine that it's not me is that I normally don't wear a tie, and this person wore a tie.
