Women are avoiding the very technology that threatens them most, as expert warns of a ‘two-tiered AI economy’ approaching

Artificial intelligence is moving even faster than many thought. In the span of three years, the world went from wearily experimenting with OpenAI’s ChatGPT to entire companies integrating Anthropic’s Claude Code into their workflows. The speed of AI’s progression, technologically and culturally, has surprised many—including Anthropic CEO Dario Amodei, who warned in a 20,000-word essay in January that society could experience catastrophic impacts within a year or two. 

But experts warn this fast-paced innovation is leaving one essential group behind: women. 

The jobs women hold are three times more likely to be automated by AI. Despite this fact, women are using AI at a rate 25% lower than men on average. This paradox is compounded by the fact that women are underrepresented in AI leadership and development, even as some of the companies with the most advanced AI adoption are led by women. 

Women are more hesitant about using AI

Leaving women out of a major technological transition could have long-term economic consequences, says workplace AI adoption strategist Mara Bolis, who warned the issue doesn’t rest with a woman’s ability to use the technology, but rather, their willingness. 

“This is not a lack of competence,” Bolis told Fortune. “This is discernment, in terms of how we want our economies and our societies to evolve.”

“I’m really worried that we’re at risk of creating a two-tiered AI economy if we don’t engage women more actively and really respect the unique skills and expertise that they bring to the field, skills that are critically important to making sure that AI evolves safely and equitably,” Bolis said. 

Bolis thinks hesitancy is a wise response to AI hype. After a stint as an economic analyst at the New York Federal Reserve, Bolis spent 11 years working on women’s economic empowerment at Oxfam. While completing a fellowship at the Harvard Kennedy School in 2023, she noticed how gender was missing from the conversation around AI policy. She founded First Prompt, an inclusive AI adoption lab that advises businesses globally on how to address and prevent inequitable AI adoption. 

Researchers at Stanford University, Harvard University, and the University of California, Berkeley found that women are less familiar with how to use AI tools and are less persistent with the technology when they use it. They are more likely to be concerned with the ethical implications of AI and about how it will affect their jobs and livelihoods. 

Women are also less certain about the benefits of AI adoption, according to Beatrice Magistro and Sophie Borwein, assistant professors of political science at Northeastern University and the University of British Columbia, respectively. The two researched how women’s risk aversion affects their skepticism toward AI’s economic benefits. 

Whether their jobs were highly complementary to AI or at risk of automation, women still perceived the technology as riskier than men did, Borwein said.  

And there’s good reason for that caution: women face a higher risk of punishment for using AI at work. A Harvard Business Review study found that female engineers are penalized more and are seen as less competent than otherwise-identical male colleagues when they produce identical AI-assisted work.

Women’s jobs will face the brunt of AI disruption 

Of the 6.1 million workers whose jobs are the most likely to be disrupted by AI and least likely to adapt, 86% are women, a Brookings analysis found. These are roles like administrative assistants, receptionists, office and legal clerks, which are positions often held by older women. Whereas men in highly AI-exposed jobs are likely to change jobs, women are most likely to completely exit the labor market rather than find new employment, Brookings found. 

“Those types of jobs that are really good, middle-class jobs. They’re well-paying jobs, they’re white-collar jobs, and they’re going to go away,” Bolis said. “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.”  

While gender disparities in AI usage persist, the gap does appear to be closing. In 2018, only 12% of machine learning engineers were women, WIRED reported. Now, 30.5% of AI professionals are women, researchers at Stanford University found. 

A September 2025 OpenAI report that analyzed 1.5 million conversations found that the gap between users with masculine and feminine names was closing. In January 2024, the company reported 37% of users had typically feminine names. By July 2025, that share had risen to 52%.

Bolis said women are in a position to find gaps with AI because they didn’t build this system. She advocates for people to approach the technology with “fierce ambivalence.”

“People think that [ambivalence] means that you don’t care, which is not what it means at all. It means holding divergent attitudes at once, which I think is very uncomfortable for people,” she said. “We need to be using AI to empower ourselves and others, while we hold the creators of this technology and the people who are setting up policies and governance to the highest possible standards to ensure that these technologies are rolled out in a way that’s safe and efficient and equitable.”

Both women and men support AI adoption when they are certain that the net effects will be positive, Magistro and Borwein’s research showed. 

“This ambivalence is not fixed. Women can lose that ambivalence if they are convinced that the net benefits are there,” Magistro said. 

This story was originally featured on Fortune.com