Why Women Face Higher Displacement Risk — and How Role Redesign Can Open New Pathways

The tech industry has spent years worrying about a talent shortage. There are not enough AI engineers, data scientists, or AI architects to go around. Companies are competing fiercely for the same narrow pool of specialists, and most of them are men.

While that war for AI talent dominates the headlines, a quieter crisis is building on the other side of the labor market. Millions of workers, disproportionately women, are in jobs that AI is already reshaping. They are not getting the same access to training, tools, or new roles that would help them make that transition.

The result is a double bind. The industry cannot find enough AI-skilled talent, while women remain the largest underutilized talent pool in the workforce. That gap between who loses work and who gains it is not random. It follows a pattern showing up in labor data across almost every major economy, and left unaddressed, it will define the gender dynamics of the workforce for the next decade.

Why women face a higher displacement risk

The headline number from the International Labour Organization (ILO) is striking: female-dominated occupations are almost twice as likely to be affected by generative AI as male-dominated ones, at 29% versus 16%. At the high-risk end, the gap is wider still. Sixteen percent of female-dominated roles fall into the most vulnerable automation categories. For male-dominated roles, that number is 3%.

The ILO report, Gen AI, Occupational Segregation and Gender Equality in the World of Work, identifies three forces driving this. Women hold the roles most likely to be automated. They are missing from the STEM fields building these tools. And AI models frequently mirror the gender biases already embedded in society.

This is not a coincidence. Women have historically been concentrated in clerical roles, administrative support, data entry, and customer service. These are exactly the functions AI handles best: routine, codifiable, and high-volume. The ILO’s research spans 88% of the countries it analyzed, and in almost all of them, women face greater exposure than men.

The exposure risk is only half the problem. The roles AI is creating are clustered in technical and strategic functions where women have historically been underrepresented. According to a 2024 study by Interface EU, globally, women make up just 22% of the AI workforce. The World Economic Forum’s 2025 Gender Parity report found that women experience a significant drop-off in the first year of STEM careers and remain underrepresented in AI engineering and leadership throughout.

Women are disproportionately concentrated in the roles being displaced, and underrepresented in the roles being created. That is not one problem. It is two problems compounding each other.

A third layer makes it worse. Randstad’s Understanding Talent Scarcity: AI and Equity report shows a 42-percentage-point gap in AI skills between men and women, at 71% versus 29%. Men are more likely to be offered AI training by employers (35% versus 27%) and more likely to have AI tools provided at work (41% versus 35%). UC Berkeley synthesised 18 studies covering 143,000 workers worldwide and found that women are approximately 20% less likely than men to use generative AI tools professionally. The gap held regardless of education level or country income.

Occupational segregation placed women in automatable roles. Underrepresentation in STEM locked them out of the roles AI is now creating. The access and training gap is preventing the transition between the two. Each layer reinforces the others.

Role redesign: what it actually means, and why most companies are getting it wrong

When organizations talk about preparing their workforce for AI, they usually mean one of two things: retraining existing employees on new tools, or replacing displaced roles with newly created technical positions. Both approaches miss the point.

Retraining is necessary but insufficient. Giving a data entry clerk a course on prompt engineering does not create a pathway. It gives her a skill set. What she needs is a destination: a specific role, with defined responsibilities, that exists in the organization and that she can credibly move into.

Replacing displaced roles with technical positions often deepens the problem. AI engineers, data scientists, and machine learning specialists require credentials and experience that few displaced workers have. They also tend to attract candidates from the same homogeneous talent pool that already dominates the tech sector. The displacement hits women. The replacement roles do not.

Genuine role redesign starts with a different question. Not what jobs can AI do, but what does human contribution look like in a world where AI handles the routine?

The answer is that distinctly human work is relational, contextual, and ethical. It is navigating ambiguity. Building trust with clients and colleagues. Making judgment calls in situations without a template. Understanding what a stakeholder actually needs, not just what they said they wanted.

The new roles emerging at this intersection carry different names depending on the sector: AI Implementation Coordinator, Technology Adoption Lead, Human-AI Liaison, Digital Ethics Officer, Change Management Specialist. What they share is a need for people who can work where technology and human complexity meet.

These roles require judgment, communication, and a deep understanding of how organizations work. They are, in other words, a direct evolution of the skills that women in today’s at-risk roles have already spent years building.

The companies getting this right are mapping the skills embedded in at-risk roles, not the job title but the actual capabilities the person has built, and identifying which of those capabilities match the roles AI is creating.

This is a talent problem, not just an equity problem

The AI talent shortage is real and getting worse. The roles being created by AI adoption require a combination of technical literacy and human judgment that is genuinely scarce. Companies are competing hard for a narrow pool of people.

Women are the largest underutilized talent reservoir in the professional workforce. The skills embedded in at-risk roles, including relationship management, operational coordination, ethical reasoning, and stakeholder communication, are exactly what the new AI-era roles require. The connection between those two facts should be obvious.

Skills-based hiring is the mechanism that makes the connection possible. Instead of filtering for credentials and linear career paths, it evaluates what someone can actually do. It opens roles to people whose capabilities developed through years in administrative and service functions, exactly the roles AI is now automating. When designed well, it does not just broaden the talent pool. It surfaces the specific kinds of experience organizations need most in an AI-augmented environment.

What it looks like when organizations get this right

There is no single model. But the organizations making meaningful progress share a recognizable set of behaviours.

They start with the skill, not the job title. Before any role is automated, they map what the person in that role can actually do, and they map it against the capabilities the organization will need going forward. The question is not whether a job can be automated. It is what the person doing that job knows, and where that knowledge fits in what is being built.

Leading organizations are moving beyond vague promises of upskilling to build pathways that are visible, specific, and actionable. Instead of a general hope for future opportunities, they provide a clear line from a current role to a defined future one, with steps, timelines, and support structures spelled out. They design training for the whole workforce, not the median employee. Programs that run after hours or require self-directed learning will systematically exclude people with caregiving responsibilities. Inclusive design means modular, schedulable, available during work hours, with the psychological safety to experiment and fail without it affecting a performance review.

This approach aligns with a fundamental shift in the workforce: the Randstad Workmonitor 2026 confirms that the traditional “career ladder” is failing, with 72% of employers now agreeing that linear career paths are outdated. In response, talent is mitigating risk by building “portfolio careers”. This new model prioritizes variety, individual agency, and security through a diverse range of experiences rather than long-term tenure in a single role.

The next 24 months will matter for a long time

Workforce transitions are not easily reversed. The patterns forming now tend to persist for years.

Organizations that act with intention can use this moment to build a more capable and more diverse workforce than they have today. Those that treat AI transformation as a technical project with a people footnote will likely emerge from it with a narrower talent base and a harder hiring problem.