New Survey Finds Women Far More Skeptical of AI Than Men, and the Gap Is Widening

Picture two coworkers sitting in the same meeting, hearing the same pitch about an AI tool that's supposed to make their jobs easier. One leaves excited. The other leaves worried about her job security, her neighborhood, and whether anyone asked what she thinks. New polling suggests that split increasingly falls along gender lines, and the gap isn't closing. It's widening.

A survey released this week by Morgan Stanley polled 2,000 consumers in late September, asking a simple question: do you feel positive or negative about AI overall? The bank then calculated net sentiment by subtracting negative responses from positive ones. Men came in at a robust plus 30%. Women landed at minus 1%.

That's not a small gap. It's a chasm, and it's been there since the bank started asking in April. Back then, women's sentiment stood at plus 6%, still well behind men, but positive. Since then, men's enthusiasm has held roughly steady while women's has slipped into negative territory entirely.

Both men and women who dislike AI tend to cite similar worries, like fears it will be used for scams or to spread misinformation. Those concerns show up in roughly equal measure across genders. But dig into the specifics and some telling differences emerge. Forty-four percent of women said they don't want a data center built near their home, compared with 36% of men. Slightly more women, 57% versus 54% of men, said AI is developing too quickly without enough safeguards. Men, interestingly, expressed more concern about data privacy specifically, at 49% compared with 39% of women.

Think of a data center the way you'd think of a new highway interchange going up next door. It promises economic activity somewhere far away, but the noise, the traffic, and the strain on local resources land squarely in your lap. That's roughly the calculus playing out for a lot of women weighing AI's costs against its benefits, and right now the costs feel closer to home.

The Jobs Problem Nobody's Fixing

Morgan Stanley's survey didn't dig into labor market data, but other research fills in that gap, and it isn't encouraging. LinkedIn data cited by Axios found women landing only a quarter or less of new high-paying AI jobs. For executive-level AI roles, the share drops even lower. Meanwhile, the jobs most likely to get automated away by AI tools tend to cluster in administrative, clerical, and customer service work, fields where women have historically made up the majority of the workforce.

That's a double bind. Women are underrepresented in the roles AI is creating and overrepresented in the roles AI is threatening. It's like watching a tide come in on one side of the beach while the other side keeps building higher ground. Bridget Griswold, CEO of the women's advocacy group LeanIn, put it bluntly to Axios: "We cannot let the AI gender gap become an entrenched chasm." Her organization has documented a related pattern inside workplaces, where women who use AI tools often face criticism or skepticism, while men doing the same thing get praised for being forward-thinking. LeanIn's research also points to evidence that AI systems themselves can carry bias against women, compounding the problem on both the input and output sides.

None of this is entirely new. Pew Research released polling earlier this year showing women were more skeptical of AI than men and used it less frequently in daily life. What's changed is the stakes. As AI moves from a novelty into something reshaping hiring, pay, and infrastructure decisions, a persistent gender gap in trust and adoption risks hardening into a gender gap in outcomes. And a workforce already grappling with a long-standing pay gap could see it stretch further if women are locked out of the highest-growth, highest-paying corner of the economy.

Age, income, and politics are also shaping how people feel about AI, and the patterns there echo the gender split in instructive ways. Adults 55 and older register net sentiment of minus 14%, compared with about plus 25% for those under 55. Lower earners, those making less than $50,000 a year, come in at plus 7%, while people earning $100,000 or more sit at plus 27%. Self-described liberals land at a flat zero, compared with plus 31% among conservatives. Across nearly every one of these divides, the people with less economic cushion or less direct benefit from AI's rollout are the ones viewing it most warily.

Michelle Weaver, head of thematic research at Morgan Stanley, told Axios that growing local pushback is already slowing things down on the ground. In some regions, data center construction has paused or stalled entirely as communities object. "It's another bottleneck for the broader AI buildout," she said. Sentiment toward data centers drops even further, the polling found, once a project is proposed near where someone actually lives. With midterm elections approaching, lawmakers in affected districts are starting to pay closer attention to constituents who feel like they're absorbing the costs of a boom they aren't sharing in.

Why It Matters

Technology adoption always comes with winners and losers, but when the divide maps so cleanly onto gender, age, and income, it stops looking like a natural market shakeout and starts looking like a policy failure. Women aren't rejecting AI out of technophobia. They're responding rationally to a landscape where the job gains are going elsewhere, the automation risk is landing on their fields, and the physical infrastructure is showing up in their neighborhoods without much say in the matter. Closing that gap will take more than better messaging. It will take genuine changes in who gets hired for AI roles, how workplaces treat women who use these tools, and how communities get consulted before a data center breaks ground nearby. Without that, the trust gap documented in this survey isn't likely to shrink. It's more likely to calcify into something far harder to undo.