Is Silicon Valley building a future for everyone, or just for half the population? We’re told Artificial Intelligence is the next industrial revolution, poised to unlock unprecedented productivity and reshape society. But the breathless coverage of Dario Amodei’s warnings about potential “catastrophic impacts” and the rapid integration of tools like Anthropic’s Claude Code into corporate workflows obscures a far more immediate and insidious risk: the widening gender gap in AI adoption. The real story here isn't the speed of AI’s advancement—it’s who’s actually using it, and who’s being left behind.
In the last three years, the narrative around AI has shifted from cautious experimentation with OpenAI’s ChatGPT to widespread implementation in businesses. Yet, despite the fact that jobs held by women are three times more likely to be automated, women are using AI at a rate 25% lower than men. This isn’t a technological hurdle; it’s a signal of something deeper. Mara Bolis, a workplace AI adoption strategist and founder of First Prompt, frames it not as a lack of competence, but as “discernment” – a considered hesitation about the direction we’re hurtling in. Bolis, who previously worked at Oxfam focusing on women’s economic empowerment, noticed this critical omission during a Harvard Kennedy School fellowship in 2023, prompting her to launch an inclusive AI adoption lab.
This hesitancy isn’t irrational. Research from Stanford University, Harvard University, and UC Berkeley reveals women are less familiar with AI tools, less persistent when using them, and more concerned about the ethical implications and potential job displacement. Beatrice Magistro and Sophie Borwein, political science professors at Northeastern and UBC respectively, have found women consistently perceive AI as riskier than men, even when their jobs are equally exposed to automation. This isn’t simply risk aversion; it’s a calculated response to a system demonstrably stacked against them. A Harvard Business Review study revealed a disturbing double standard: female engineers are penalized for using AI assistance, perceived as less competent than their male counterparts producing identical work.
Drawn from Fortune.
The economic stakes are enormous. A Brookings analysis found that 86% of the 6.1 million workers most vulnerable to AI-driven job disruption are women – primarily in administrative and clerical roles, often held by older women. Unlike men in similar positions who are more likely to transition to new jobs, women are far more likely to exit the labor market entirely. Bolis paints a stark picture: “Those types of jobs…they’re going to fall into less well paid, less secure work as that entire sector falls away, unless we focus intentionally on creating policies and programs that help them weather this change.” This isn’t about retraining for coding bootcamps; it’s about protecting livelihoods and preventing a massive rollback of economic progress for women.
However, the situation isn’t entirely bleak. While the gender gap in AI engineering was a dismal 12% in 2018, it’s now climbed to 30.5%, according to Stanford researchers. More encouragingly, data from OpenAI shows a significant shift in user demographics. In January 2024, 37% of users had typically feminine names, rising to 52% by July 2025. This suggests that as AI becomes more accessible and integrated into everyday tools, the initial hesitancy is beginning to wane. Bolis advocates for a mindset of “fierce ambivalence” – embracing the potential of AI while simultaneously holding developers and policymakers accountable for equitable implementation. She argues that women, as relative outsiders to the initial development of these systems, are uniquely positioned to identify critical flaws and advocate for responsible innovation.
The key, according to Magistro, is demonstrating clear net benefits. Both men and women support AI adoption when convinced it will lead to positive outcomes. But convincing women requires more than marketing hype; it demands addressing the systemic biases embedded within AI systems and actively mitigating the risks of job displacement and unfair evaluation. The closing of the user gap at OpenAI is a positive sign, but it’s a lagging indicator.
Looking ahead, watch for the rise of “AI fluency” certifications specifically designed to address the concerns of women entering the workforce. These won’t be about coding; they’ll be about understanding how AI impacts job security, recognizing algorithmic bias, and leveraging AI tools to enhance – not replace – existing skills. If these certifications don’t emerge, and if companies don’t proactively address the gender imbalance in AI adoption, we’ll see a widening economic divide, not a technological revolution. The question isn’t if AI will change the world, but whose world it will be.
