In the 20 years I’ve been a TV producer, media executive, and speaking coach, I cannot count the number of times I’ve seen people literally cringe upon hearing their own voice. Sometimes it’s a slight scrunch of the nose and sometimes it’s an audible …
Stores are cutting back on self-checkout lanes.
According to an April survey from Toast, a point-of-sale and restaurant management platform, approximately 36% of 340 small and midsize retail operators said they used self-checkout lanes this year, d…
The debate about AI and work often focuses on the jobs that will disappear, which tasks will be automated, and which skills and qualities will be valued in the future. It is more seldom that I read articles that analyze the impact of AI on exhaustion and pleasure at work. And those questions are usually asked about “workers” in general, as if the answers were the same for men and women. There are many reasons why they are not. Because women and men are concentrated in different occupations, perform different tasks within the same organizations, and are judged differently for the same work, AI affects women’s work in very specific ways.
Is the overall effect positive or negative? It isn’t easy to say, so I made an old-fashioned list of pros and cons, the kind people draw up when deciding whether to change jobs or leave a relationship. I ended up with more negatives than positives: five reasons to be hopeful and eight reasons to worry.
On the hopeful side
#1. Today’s job growth is happening mostly in women’s jobs, and they aren’t very much exposed to AI. In 2025, health care and social assistance accounted for almost all net job creation in the US, about 405,000 jobs in health care and 308,000 in social assistance. If it weren’t for those two sectors, the US would have lost jobs last year. The BLS anticipates that this will continue to be the case in the coming decade, because of population ageing. We will need a lot of caregivers in the future, and LLMs can’t take their place. That protects women’s employment, but it does little for their pay: care work remains badly paid, and in the US its growth also depends on public money that hinges on political choices. The coming midterms will say a lot about where things are heading.
#2. AI could make “greedy jobs” less hostile to women. A big part of the pay gap is attributable to the unequal distribution of the highest-paying jobs. The so-called “greedy jobs” reward presenteeism and excessive workloads, which penalizes everyone who can’t give work all of their time. In law, finance, consulting or medicine, value is tied to being the one person who knows the client or the case. If AI makes expertise portable, making a client’s or a case’s history immediately accessible to other colleagues, then it could reduce the premium on being irreplaceable. This is what digital records did to pharmacy, which went from a male-dominated, unequal profession in the 1970s to an equal one in the US. Of course, we’ll still need employers to decide to redesign those jobs, and that can never be guaranteed.
#3. Hurrah, some of the most tedious office chores are now being automated. Writing meeting minutes, taking notes, organising schedules, drafting follow-up emails: most of these “non-promotable tasks” are important but do nothing to advance the career of those who do them. And they fall predominantly on women’s shoulders (it shouldn’t come as a surprise: they’re exactly like household chores). The time spent on them could be spent on more visible tasks that can earn you a promotion. So AI tools that automate them are more than welcome! Likewise, there’s less paperwork in care jobs. Nurses and doctors are worn down by paperwork. Using AI tools can help them focus more on patients, which happens to be the part of the job they like best.
#4. The barriers to entrepreneurship and to building with tech are lower now. Women who start companies have always had to do more with less. In the US, startups founded only by women still get roughly 2% of venture capital (and this share isn’t budging). AI makes it possible to start small and still go far. On your own, you can accomplish so much more with an army of “agents”! You can create a website with very little money, draft contracts, answer customers… According to Carta, the share of new startups with a single founder rose from about 24% in 2019 to 36% in the first half of 2025. AI-assisted coding also enables people to develop digital tools without knowing how to code, which is remarkable in a field where women are in the minority.
#5. Relational skills are gaining value. As AI takes over the codifiable part of cognitive work, the value of what it doesn’t do so well could go up: coordinating, caring, exercising judgment in context. A lot of women have acquired these skills through decades of being expected to look after others and manage their emotions. If AI does raise the value of those skills, the market may have to pay more for them and stop seeing them as “naturally” feminine and therefore of little value.
On the worrisome side
#1. The most exposed white-collar jobs are disproportionately held by women. In rich countries, according to the International Labour Organization, 9.6% of women’s jobs fall into the highest category of exposure to generative AI, compared with 3.5% of men’s jobs. The reason is that women are concentrated in clerical jobs: assistants, secretaries, payroll and data entry clerks. Not all these jobs will disappear. But we already have an idea of how they can change. Translation is a mostly female profession (around seven in ten translators and interpreters in the US are women). A big share of translators have lost work to generative AI (and others have changed professions already). Many of those who still have work now spend their days editing machine output, for less money.
#2. The first rungs of the ladder are disappearing. For generations, junior administrative and clerical jobs were a route into employment, and upward mobility, for women without higher education. A secretary could become an office manager, then an executive assistant, then an operations manager. If AI eliminates most of these entry points, the career ladder gets cut at the bottom, and that’s bad news for women.
#3. Women who use AI pay a competence penalty. In a study of more than 1,000 engineers reported by Harvard Business Review, identical code was judged differently depending on who was known to have written it. Women who used AI saw their perceived competence drop by 13%, compared with 6% for men, and by 26% when the reviewers were men who did not use AI themselves. This is a digital version of the Matilda effect: when a woman uses the machine, the credit goes to the machine; when a man uses it, his judgment in choosing to use it is valued. Women are caught in a double bind: they are judged less competent if they use AI and less up to date if they don’t.
#4. Women are largely absent from one of today’s most lucrative industries. AI is already one of the best-paid segments of the market. According to LinkedIn, women accounted for only 26% of AI hires in 2025, compared with half of all hires in other sectors, and they hold only 13% of C-suite roles at AI companies.
#5. The tools are built mostly by men, and it shows. LLMs learn from data that reflects a world where men have long held the best jobs and earned the highest salaries. German researchers at the Technical University of Würzburg-Schweinfurt asked chatbots for salary negotiation advice using profiles that were identical except for gender. Women were told to aim lower. In one case, the model suggested $280,000 for a woman and $400,000 for a man. Years earlier, Amazon had to drop an experimental recruiting tool after discovering it penalised CVs containing the word “women”, as in “women’s chess club captain.” The system had been trained on ten years of applications, most of them from men. In hiring, evaluation and career advice, the same biases get repeated at scale.
#6. New office chores replace the old ones. Someone still has to read the automated minutes, improve the AI-generated drafts and catch the made-up figures and other hallucinations before they go out. AI slop takes time to process. I suspect a lot of this “cleanup” lands where office chores have always landed. If that’s right, some of the time saved by automating the old chores gets eaten by the new ones.
#7. The jobs AI can’t do are the ones it watches most closely. Call centers and home care are good examples. In the US, 85% of home care workers are women, and since 2020 those paid through Medicaid have had to log every visit in an app that tracks them by GPS and checks their identity. Call centre and retail workers know the same timed tasks and automated schedules. These women keep their jobs, but they lose more and more control over how they do them.
#8. Deepfakes make visibility more dangerous for women. Nearly all deepfake videos online are pornographic, and 99% of the people targeted in them are women. Public life makes it worse: in the US Congress, women were 70 times more likely than men to be targeted. Journalists, politicians, founders, anyone whose career depends on being seen now faces a risk men rarely do. Many women are likely to respond by stepping back (and who could blame them?) After facing online abuse, 38% of women journalists surveyed by UNESCO said they had lowered their public profile.
On balance
My list clearly has more negatives than positives, which in itself doesn’t mean much. Counting items doesn’t tell you how much each one weighs. Besides, the list is far from comprehensive. You may have other arguments on the credit or the debit side…
Still, having drawn up both columns, I would say the balance is negative for women’s work, at least for now. The main reason is a difference in timing. Most of the negatives are already happening: women hold the most exposed jobs, the first rungs of the career ladder are disappearing, women who use AI are judged more harshly, they are a small minority in the sector where the money is, the tools carry biases, the jobs AI can’t do are increasingly monitored, and deepfakes make public visibility riskier. Most of the positives, on the other hand, depend on conditions that haven’t yet been met: greedy jobs will only loosen if firms deliberately redesign work, the time freed from office chores and paperwork can easily be swallowed by new tasks, relational skills will only pay if employers finally choose to recognise them…
The balance could still shift, of course. What happens next will depend on choices made inside companies, about job design, evaluation and hiring in AI, and on choices made at the ballot box, about how much public money goes to care.
I don’t consciously think about my first job when I meet with leaders in the electronics industry to talk about trade policy. But my young days in an Illinois cornfield will always shape how I view, and do, work.
My hometown in western Illinois had…
For decades now, American workers have adhered to an unspoken deal that, in order to be successful professionally, some sacrifices need to be made. Whether it be checking work emails after dinner, taking work calls while on vacation, or being willing …
Advertising is made for the young by the young.
That’s conventional wisdom backed up by the Bureau of Labor Statistics, which in 2025 reported the median age of ad industry workers was 39, compared to 42 for the broader workforce. Meanwhile, …
Below, June Cohen shares five key insights from her new book, The Great Remembering: Powerful Ancient Practices―And the New Science Bringing Them Back.
Cohen is a pioneer in media and technology, best known as the co-creator of TED Talks, …
Would you be as proud to tell someone your child is an electrician as you would be to say they are a business major?
We rely on these workers to build our infrastructure, power our industries, and keep essential systems running. Yet we still …
Plenty of CEOs are bullish on AI and have urged their employees to embrace it enthusiastically. A more unusual approach is to deride your workers as losers for not using AI enough.
A new investigation this week from 404 Media shows how one CE…
Companies around the world are increasingly citing artificial intelligence when slashing jobs, fueling anxiety that a lot more workers may lose their livelihoods to machines.Microsoft, Barclays Bank and supermarket chain Kroger are among the latest en…
