Copyright Project Syndicate
LONDON – In 1986, “Anne” – a data entry clerk – watches an IBM PC land on her desk. Within a year, her job is gone. Four decades later, “Natalie” – a social-media manager – looks on as ChatGPT drafts the posts she once wrote. But her exit may come even faster than Anne’s. In July, a new report from Microsoft researchers made headlines by listing the 40 occupations most at risk of being replaced by AI. It included sales reps, translators, proofreaders, and other knowledge jobs, pointing to a looming white-collar employment apocalypse. The report’s authors, and the subsequent news coverage, seem to have overlooked, however, something crucial: the coming disruption will not be gender-neutral. According to the US Bureau of Labor Statistics (BLS), women comprise the majority of workers in around 60% of the occupations listed. While AI is out to eat everyone’s lunch, it’s women’s lunches the technology looks likely to gobble up first, and fastest. Just as the advent of computers in the 1980s displaced legions of secretaries and data-entry clerks – positions held mainly by women – so, too, is this latest wave of automation likely to have a disproportionate impact on female workers. According to recent research by the International Labour Organization, women’s jobs in high-income countries are roughly three times more likely to be automated than men’s. The computer revolution serves as a cautionary tale. Many of the women who lost their jobs in the 1980s because of it never recovered, either finding lower-paid work (primarily in the service and care sectors) after protracted periods of unemployment or leaving the workforce altogether. When the BLS tracked the outcomes of workers who were displaced during this period, the findings were stark: women were more than twice as likely as men to have subsequently dropped out of the labor market. Given that women are already at an economic disadvantage relative to men – they earn less, own less, and retire with less – policymakers need to prepare for AI hitting women’s jobs the hardest and put in place policies to mitigate the impact. When crafting their response, they would do well to note too that not all secretaries, data-entry clerks, and typists fared equally badly in the 1980s: the women who managed to adapt to the technology and acquire relevant new skills had better outcomes. Set aside, for now, the question of whether the concept of “upskilling” is redundant in an era in which AI is expected to surpass human intelligence, and instead assume that there will be a transitional phase in which workers with AI-related skills do better than those without them. PwC’s 2025 Global AI Jobs Barometer found that workers with AI skills command a 56% wage premium, a dramatic increase from the 25% wage premium reported the year before. What this suggests is that if we are to prevent female workers from becoming AI’s most immediate collateral damage we must ensure that they are fully up to speed on the new technology – or at least as much as their male counterparts. But while roughly equal numbers of women and men now use ChatGPT for personal tasks, a clear gender divide has emerged in the workplace. A recent survey of US workers revealed that while 36% of men use generative AI daily on the job, only 25% of women do. It also reports that 47% of men say they are confident using the technology at work, compared to 39% of women. This gap likely reflects the fact that women are more concerned about the increasing use of AI than men– a healthy skepticism we should all retain. But another reason is that companies are investing more in AI upskilling their male employees than they are their female ones. In a global survey of 12,000 professionals conducted this year by Randstad, 41% of men said they had been provided with AI access by their employer, compared to 35% of women, while 38% of men said they had been offered opportunities to build AI skills, compared to 33% of women. Using the technology less – and being given fewer chances to use it – is a toxic combination for female employees, especially as firms increasingly cite “AI fluency” when deciding who to retain and promote. Failure to address this problem could also expose their employers to legal risk. In the United Kingdom, workplace policies that systematically disadvantage women – and offering fewer opportunities for AI upskilling may well fall into this category – could constitute indirect sex discrimination under the Equality Act 2010. This is true even if a firm did not set out to discriminate. Under this law (and similar legislation in other countries), what matters is impact, not intent. Business leaders should thus be asking themselves basic questions. Who is getting access to AI tools? Who is being invited to participate in AI pilots and initiatives? Who is receiving AI training? Governments around the world seem wholly unprepared for the potential jobs Armageddon that AI could trigger, especially as it affects women. As policymakers develop AI risk-mitigation strategies, it is imperative that gender concerns be firmly on the agenda – and not just for ethical reasons. At a time when political polarization is increasing and traditional parties are losing ground, winning over female voters will be pivotal. As such, ensuring that women don’t bear the brunt of AI-induced job displacement, and addressing other AI-related gender disparities as well, is not only the right thing for governments to do. It is also highly pragmatic. After all, there are a lot of Natalies out there.
