In higher education today, students are being encouraged to pursue careers that are “safe” from AI, making entering the job market after university all the more daunting. But what does it mean for a job to be safe from AI, and who is occupying those jobs that are safe versus those that are vulnerable to AI? One of the most common conversations surrounding AI is the fear that it is taking jobs away from hardworking humans, and thus fundamentally altering the job market in unprecedented ways. International Labour Organization (ILO) research suggests that “GenAI is set to affect women’s jobs more than men’s, with female-dominated occupations almost twice as likely to be exposed to the technology” because the jobs that are most easily automated are often ones traditionally held by women. Often overlooked are the systemic reasons for this discrepancy, which must be addressed before any attempt to rectify this disproportionate insecurity can be made. Women hold more of the positions that are easily automated, and thus vulnerable to replacement by AI, largely due to patriarchal norms characterizing women as naturally suited to that work. AI can reasonably be expected to exacerbate structural inequalities that already exist between men and women in the job market, and the sooner we recognize this, the sooner we can stop it.
What makes a job vulnerable to replacement by AI is how easily it can be automated. How hard a job is to automate depends on the extent to which it relies on human relationships, manual labour, and creativity—skills that cannot (yet) be replicated using AI without sacrificing quality. On the other hand, jobs that do not rely on these skills, but rather on repetitive systems, like clerical work, data entry, or customer service, are far easier to program an artificial system to do, posing the threat of replacement to people who currently occupy these positions. In countries such as the United States, the Philippines, the UK, Ecuador, and Botswana, data show these fields to be dominated by women.
Largely due to deep-seated patriarchal norms, women still dominate industries like customer service and administrative support, even as job markets have become more open than they once were. These norms cast women as more personable and caring, keeping with the stereotype of women as natural caregivers. In customer service, for example, friendliness, gentleness, and compassion are highly prized traits, because they make customers feel calm, safe, and appreciated, which is likely to increase sales. In administrative support work, a welcoming or caring nature is also valued. Additionally, these roles are often seen as subordinate to, or in service of, a company’s higher-ups, which is thought to be a position more suitable to women.
These jobs, in addition to being seen as suitable for people with stereotypically feminine traits, are also often entry-level positions. There is far less talk of managers or CEOs being replaced by AI, both jobs that are shown to be more male-dominated. Entry-level positions are far more vulnerable to job loss, and also have a tendency to be female-dominated. And this fate is not one that is easy for women to escape. In one study, women were shown to be 14% less likely to be promoted than their male colleagues. They were also shown to occupy less than 15% of CEO positions at Fortune 500 Companies. This gap is also largely attributable to societal stereotypes of men as more assertive and dominant, traits seen as desirable for a managerial or superior position, as this MIT paper suggests. Women are concentrated in the jobs most vulnerable to AI, and are also kept out of the senior roles that are safer from it.
Policies preventing AI job loss are new and still thin, but a few initial policy attempts exist. In Hangzhou, China, a court ruled that replacement with AI is not sufficient justification to fire employees because utilizing AI is a voluntary decision made by a company, not an external circumstance they could not control. This brings up an important point—though the use of AI is often a profitable choice, it is still a deliberate one, so the consequences and implications of doing so fall on the company to manage and rectify. The AI-Related Job Impacts Clarity Act (bill S. 3108) was proposed in the U.S. senate in late 2025. This bill would require companies to submit a quarterly report on the impacts of AI on jobs, but it has not yet made it out of committee.
Policy about AI exists, and policy about AI taking jobs exists. None of these measures, however, aims to ensure that AI-driven layoffs affect people equitably, so nothing prevents them from falling disproportionately on women, as they are anticipated to. Recognizing this projected inequality early is an important step towards preventing further structural inequality in the job market.
Historically, we have seen what happens when women are less involved in the workforce: they are less involved in the economy, which leads to them being less involved in and represented in politics, which puts issues pertaining to them lower on the list of government priorities. In order for women to be represented, protected, and involved, their jobs must be secure and protected to the same extent as anyone else. AI is dangerous in more ways than one, and by calling out its reinforcement of systemic inequalities, we can create real change and get AI moving in a more helpful direction.
Edited by Stella Fox
Featured Image: Bill Branson (photographer), Public domain, via Wikimedia Commons
