The AI Gender Gap Is Narrowing. But the Hard Work Is Still Ahead.
Women have caught up with men when it comes to trying AI, but research shows there’s still a big gap when it comes to confidence, trust and perceived value.
Some good news for anyone concerned that the gender AI-use gap is exacerbating inequality: Research recently conducted by the Pew Research Center found that a similar number of men and women now say that they use AI chatbots, marking a considerable change from two years ago when men were 11 percentage points more likely than women to say they use them.
The survey of more than 5,800 U.S. adults, conducted in February of this year, found that women have largely caught up with men when it comes to trying AI. That's an encouraging milestone. If AI is becoming an increasingly important tool for work, learning and everyday life, a narrowing adoption gap matters.
Dig a little deeper and the picture becomes more nuanced. Men are still significantly more likely to be frequent users. More than a quarter — 27% — say they use chatbots daily, compared with one in five women, or 20%. Men are also more likely to use a wider range of models, including Copilot, Gemini, Claude and Grok, while ChatGPT is the only platform used equally by both sexes.
The confidence gap is alive and well, too. Pew found that men are consistently more optimistic about what AI can do, both for themselves and for society. They are more likely to say AI boosts their productivity, creativity and knowledge, while women remain markedly more skeptical about its long-term impact.
Kathleen Davis
Around one-third of women believe AI will negatively affect them personally over the next two decades, compared with just 17% who expect a positive impact. Men, by contrast, are almost evenly split between optimism and pessimism. Women are also considerably more likely to say AI is advancing too quickly.
None of this should be interpreted as a simple call for women to simply embrace AI more enthusiastically. As Ruchika T. Malhotra recently wrote for The Persistent, we need to stop telling women what to do when it comes to AI. Given the technology's well-documented problems — from bias and privacy concerns to hallucinations and labour disruption — women’s caution and anxiety (as documented in this fascinating paper from 2025) are not only understandable but entirely rational.
We should also remember that research has explicitly shown that women are more harshly judged for using generative AI on the job. One study on this, found that female engineers who used AI to generate code were rated 9% less competent than their male peers, despite evaluators reviewing identical outputs.
Instead, what this new Pew research should do is serve as yet another important reminder of the fact that we all need to do a better job of navigating this new era of AI.
We need chatbots that aren’t sexist or racist. And in order for such chatbots to exist, we must increase representation of historically under-represented individuals in the companies that are building these AIs. We must also have more education and stronger regulation around data sharing and privacy.
Many organizations have yet to create the conditions and tools that enable everyone to use these new technologies confidently and critically. Too many workplaces still expect people to just figure AI out for themselves: to wade into the great unknown, replete with threats, opportunities and risks.
Closing the adoption gap is progress. Closing the gap in trust, confidence and perceived value will be much harder — but also much more important. If AI is going to shape the future of work, ensuring everyone feels equally equipped to use it is mission critical.
Emma Haslett
