LinkedIn’s recently published Economic Graph shows that women account for just over a quarter (26%) of new hires into AI roles. It’s a sad, yet unsurprising headline. And it’s sad that it is unsurprising.
Here’s a question. Who gets to shape the technology that is reorganising the way we live and work more and more each day? The technology that will shape every space and place on the planet. The follow-up question would be: but what happens when half of the population is largely absent from that process?
“It’s worth looking back and admiring our real progress. We’re in a far better place than 50 or 100 years ago. But it’s also a day that surfaces recent conversations that reveal the same challenges still sitting underneath. The devil’s-advocate argument that hiring a woman is ‘riskier’ because she might go on maternity leave. The call to ‘go back to meritocracy,’ as if striving for equity means something other than that exact thing. The idea that women being a minority in tech but a majority in education and HR is simply about individual interest – nothing to do with the systems around us.
“That’s what strikes me. Zoom into the details, and it’s the same arguments, the same biases, the same reasons to brush off the idea that how we work together deserves the same deliberate effort as what we work on,” says Liina Adov, People & Culture Community Manager and personal coach at Pipedrive.
A love-hate relationship with AI
The way I see it, women’s relationship with AI currently looks a lot like a love-hate relationship.
On one side, there are women who want to embrace AI and have a real say in how it works. They want to shape it. They want to shape not just who it represents but also how it represents them. They see the future it could represent.
On the other side, there are reports of women using AI tools less than men, often out of fear of being seen as less competent, or because their workplaces haven’t set out clear guidelines on how AI should – and shouldn’t – be used. There is also a stigma attached to women that if they use AI, they can be perceived as less capable in their roles rather than more capable and efficient.
And then there are the headlines warning that AI is displacing women’s jobs faster than men’s, with female-dominated roles among those that are the most exposed to automation.
When it seems that every avenue has a gate that turns women onto another uneven path, is it any wonder there can sometimes be uncertainty on the best way forward?
LinkedIn’s numbers show a crumbling career path
According to the report, in the US in 2025, women made up just 26% of new hires into AI roles, compared with 50% of new hires into non-AI occupations. But even then, that gap isn’t evenly spread. It’s widest in the best-paid, most senior positions. Women accounted for only 20% of new hires into Head of AI roles, 26% into Director of AI roles, and 18% into Member of Technical Staff positions – which is one of the fastest-growing AI jobs that carries with it a median listed salary of $220,000+.
If we look to the other end of the scale, the lowest-paid AI role is that of data annotation. It carries with it a median salary of around $51,000, and it is also the most evenly split between males and females.
But this isn’t just a US problem. Across 27 countries, women hold just 13% of C-suite roles at AI companies, compared with 45% of the workforce in equivalent non-AI companies. Even at director level and above, women’s share drops from around 38% generally to just 31% within AI leadership specifically.
Women are equally represented in the lowest-paid, least powerful corner of AI. While the roles that set direction, budget, and priorities stay overwhelmingly male.
This is bigger than a hiring statistic
These latest findings matter more than a job title. AI is being built to make decisions about hiring, healthcare, lending, policy, creative work, etc. The output of its input will affect everyone. If the people building and directing it skew heavily toward one half of the population, or a collective way of thinking, the tools, which are there to carry out instructions, risk carrying that imbalance forward at scale – where it will be encoded into the products that will be used by everyone.
“AI is not neutral. Left alone, it doesn’t erase old inequality, it learns and repeats it, faster,” says Adov.
“A world in which fewer women are involved in the production and regulation of AI is also one that threatens to produce models that fail to reflect – and support – the needs of half of the world’s population, while a world in which fewer women possess AI skills is one in which women are disproportionately excluded from the economic and professional opportunities that the technology is creating,” says Marni Baker Stein, Chief Content Officer, Coursera.
According to experts, this imbalance points to two causes in particular. The long-standing gender gap in STEM education and a leaky talent pipeline that loses women at almost every stage between entry-level hiring and the C-suite. Neither of these is new. Both are enormous concerns. Yet, neither has an easy fix.
The education pipeline – are headlines helpful?
When we consider the narrative that there is a long-standing gap in gender representation in STEM subjects, I can’t help but ponder the exam-results narrative that has hit the UK headlines these past couple of weeks. Reports comparing girls’ and boys’ A-Level and GCSE results are making it a girls vs boys situation. Framing that story as a competition manufactures rivalry, and that helps no one. Isn’t it more useful to ask why grades are slipping at all and what that could mean for the pipeline of future talent and leaders, especially in STEM, for everyone?
What could actually get the needle moving?
There is no fast and magic solution, but perhaps a good place to start is upskilling the people already in the workforce. AI is going to be part of most jobs, not just “AI jobs”. So giving existing employees proper training and clear guidelines on how to use these tools could well go some way toward closing the confidence and competence gap.
“Ensuring that women around the world are provided with opportunities to gain AI literacy is one of the most important upskilling priorities of our time,” says Baker Stein.
However, there are signs of progress. On Coursera, female learners accounted for 36% of AI enrolments in 2025—up from 32% in 2024,according to Baker Stein. “We must accelerate efforts towards parity, as we are already seeing employers foreground AI literacy in setting out hiring and promotion criteria. One recent Coursera survey found that 72% of leaders expect new hires, regardless of their specific role, to understand how generative AI could be applied to their work tasks.”
“Closing this gap therefore means ensuring women not only have the skills to navigate an AI economy, but a central role in shaping it. Encouragingly, research into learning science identifies tried-and-tested techniques to drive female participation in these domains: mentorship, clear job-relevance, and ensuring that learning is flexible and accessible,” says Baker Stein.
One of the selling points of AI is that it is efficient and democratises access. It exudes the promise that it lowers barriers rather than raising new ones. But that only works if the people designing and deploying it actually reflect the population using it. Companies asking the world to adopt AI at scale need to reckon with the fact that half of that world is female, and women have differing perspectives and priorities. Those differing views don’t just deserve a seat at the table (even though they certainly do deserve one) – they need one, especially if the technology is going to work for everyone it touches.
“We’re at a point of huge disruption and opportunity right now, but only if we’re deliberate about it … The opportunity is to purposely design for something better, being mindful of who gets access to training, who leads, and whose judgment shapes the tools we build. More voices, more varied experience: fewer blind spots, better decisions, stronger products. That takes effort now, and it pays off in what we build,” says Adov.
“It is essential that we deploy these techniques at scale, empowering all women to gain the critical skills they need to thrive in the GenAI economy. This Women’s Equality Day, the progress we are seeing gives us reason for optimism, but also a reminder that we must do more, more intentionally. It is our collective responsibility to ensure that all women can not only participate in the GenAI economy, but help shape its future,” says Baker Stein.
