Many workers are worried that their jobs will be replaced by artificial intelligence, and a growing body of research suggests that women have the most cause for concern.
A recent report from the International Monetary Fund found that in advanced economies such as the United States, about 60% of jobs may be impacted by AI. Of those exposed, researchers estimate that half will become more productive thanks to AI, and for the other half, “AI applications may execute key tasks currently performed by humans, which could lower labor demand, leading to lower wages and reduced hiring. In the most extreme cases, some of these jobs may disappear.”
According to the IMF, women are at most risk for the latter fate. “The result can be interpreted to mean that women face both greater risks and greater opportunities,” reads the report.
LinkedIn has also recently found that women are more likely to be replaced by AI. An analysis of LinkedIn users’ data around the world indicates that 56% of men, but just 46% of women, will have their jobs augmented by AI. And while 43% of men will have their jobs disrupted by AI, 57% of women will.
In the U.S. specifically, LinkedIn has found that more than a third of women work in occupations that can be disrupted by AI, compared to roughly a quarter of men.
The job search giant suggests that occupational segregation is partially to blame for these differences. For instance, jobs in which women make up the majority of workers—such as medical administrative assistants (91% women), office managers (88%), and legal assistants (87%)—may be more likely to be replaced by generative AI. Meanwhile, jobs in which men make up the majority of workers—such as electrical engineers (94% men), mechanical engineers (89%), and computer network technicians (88%)—may be more likely to benefit from the use of generative AI.
LinkedIn has found that this trend appears to extend beyond the U.S. and across a wide range of industries. Fast Company spoke with Sarah Steinberg, LinkedIn’s head of global public policy, to discuss how AI will impact women—and what can be done to improve the future of work for those affected by this disruption.
This interview has been edited for length and clarity.
LinkedIn found that in the U.S., more than a third of women work in occupations that could be disrupted by AI, compared to 25.5% of men. How did you come to this estimation?
We mapped the skills that are associated with different jobs against the skills that generative AI can do. Then we broke them into these three quadrants. First, there are the jobs that are fairly insulated—generative AI can’t do a large proportion of the skills that are associated with those jobs. Next, there are those jobs that can be augmented [by AI]; they require more uniquely human skills as well as skills that can potentially be enhanced with generative AI. And then this disrupted bucket includes jobs that are more likely to have a greater share of skills that can be performed by generative AI.
Obviously, this is a relatively new technology and this is going to play out over a number of years.
Is there something about these jobs that are disproportionately held by women that makes them inherently more replaceable by AI? Or is it that we’re designing AI to replace the work of women?
It’s an interesting question. This is a very transformative moment, which on the one hand can generate a lot of anxiety—even feelings of fear—about what this means. But on the other hand, these sort of moments of great transformation also present opportunities to shape a future where women are equally represented in the economy.
What we see is that, yes, there is a rise in the importance of both AI engineering skills as more companies are trying to build these products, as well as AI literacy skills as more people are [learning to utilize tools like] ChatGPT and GitHub Copilot so that they can use AI to augment their jobs. At the same time, though, on LinkedIn we’re seeing a rise in demand for soft skills and people skills like communication.
When we do our executives survey we see overwhelmingly that executives say some soft skills are more important than any AI skills. And actually, women have more of these soft skills. We see that gap widening over the last eight years instead of decreasing.
What that points to is an opportunity. If we’re thoughtful about how we’re leveraging things like skills-based hiring (to hire people on the basis of the skills they bring to the job, instead of just their degree or their previous fancy job title) and intentionally upskilling so that we’re getting women those AI literacy skills that they need to augment their roles, this is the place where potentially women have an advantage. Because women actually have more of those uniquely human skills like communication, collaboration, teamwork, and conflict negotiation—and all those sorts of things that we can’t see AI doing very well.
How optimistic are you that employers will value these soft skills? Because I often see employers who are forced to shell out money for hard technical skills. But I less often see the same play out for soft skills.
At LinkedIn, a lot of our research doesn’t have a great angle on pay, but what we can say is this is definitely a moment to be intentional about how we’re developing those soft skills.
Before I came to LinkedIn, I was in corporate philanthropy for a long time, working on supporting education and training programs and trying to give people the skills they need to move into well-paying jobs that don’t necessarily require a college degree. And I think I was guilty—and a lot of us were guilty in those days, in the 2010s—of focusing very narrowly on those technical skills.
It was the era of coding boot camps and the idea that technical skills were the most important thing you needed to get into the jobs of the future and digital jobs. We weren’t thinking as much about those soft skills. And what we’re seeing now is that in a world in which AI can do a lot of these more rote, routine tasks, that really elevates the importance of soft skills.
What are some thoughtful ways we can upskill workers to make sure that the future of work is more gender equitable?
It’s important that we close the gender gap in terms of AI engineering. There is a massive gender gap in terms of who’s designing these tools and products and services, and that needs to be remedied, because you can’t design products that are going to meet the needs of the entire population of the world when only men are designing them.
But we don’t necessarily need everyone to be developing AI engineering skills. We don’t need everyone to understand computer vision and natural language processing. But we are moving to a place where everyone is going to need AI literacy skills. You’re going to have to use Copilot, and ChatGPT, and all the tools that are going to help you do your job better and more efficiently. So I think focusing on those AI literacy skills [is particularly important] for women.
When it comes to soft and people skills, we need to be really intentional about developing those, because I think it’s been easy to sort of err on the side of training around technical skills because it seems more concrete; you can evaluate them pretty easily. But the softer skills are just as important, and they’re a little trickier to deliberately cultivate and evaluate.
How are different regions of the world being impacted by the rise in demand for AI skills?
We see AI skills rising from all over the world, in every country in the world, in every industry. Obviously, there are some countries and industries that are leading, but we’re seeing tremendous growth across the board. The thing that struck me from the gender research was that we see that gap, in terms of women being overrepresented in the more AI-replicable jobs than men, in just about every country in the world. It really drives home the point around this being driven by so much occupational segregation, and how much that occupational segregation exists in every country in the world.
We looked at 74 countries and in almost every single one of them women were overrepresented in AI-replicable jobs.
Are there any organizations that you think are doing a really good job in terms of upskilling women, or just workers in general, for the AI revolution?
No. We did an executive survey and the thing that jumped out to me was that almost everyone surveyed was like, “Yes, we see the opportunity for generative AI to increase our productivity and revenues.” But then when you went down to the level of are you actually thinking about how you would reorganize job roles or restructure, only 4% of executives surveyed said they had gotten to that level. So I think it is pretty early days in figuring this out, which, again, gives us the opportunity to take this moment—and hopefully leaders can be really thoughtful about it—and figure out how we support women to get that equal foothold [without] replicating the inequities that have been built into the system thus far.
In terms of upskilling, are there any approaches that are especially effective? Partnerships with public universities, or funds given directly to companies to upskill their workers, or individual-level investments in services like LinkedIn Learning?
All of the above. For so many years, we’ve gotten stuck in this model where you front-load education at the beginning of your career and expect it to last for an entire career and get you through and keep you advancing. I don’t know if that was ever the optimal model, but certainly in the world today it’s not.
One of the things we documented at LinkedIn is, even before the introduction of generative AI, if you look at the same job today, and in 2015, the skills required for that job have changed by 25%. That’s the same job. The same job title. So now with the advent of generative AI, we expect that those skills will change by 68% by 2030. Think about that: 2030, same job title as 2015, more than two-thirds of the skills are going to be different. So what that points to is the need to be much more agile and flexible around how you’re intentionally upskilling throughout the course of a person’s career.
I think all of those things are going to be important, but focusing on being able to meet people where they are, help them identify what are the skills they need to move into different roles, and then having a system that recognizes and values those skills so that we’re hiring people on the basis of those skills, and not the degree that they got however long ago.
One fear that a lot of people share is that AI is going to lead to sustained, elevated unemployment. Given the data you’ve shared, my fear is that if and when we hit that inflection point, women are going to be disproportionately unemployed. What do you make of these fears?
This is a transformative moment. We do have to be purposeful about how we’re creating opportunities for women to achieve equality in this new era through upskilling and skills-based hiring, education, and training. I will say that from the outset.
But we don’t know how these things will shake out, and at the end of the day, unless there is a limit to human desire, there’s probably not a limit to economic growth and innovation. We expect that [AI] will change jobs, but will it lead to a world in which we’re all permanently unemployed? Probably not likely—unless there’s sort of like a fundamental shift in human desire.
