Last week, I asked Sam Altman a question that has defined my career. I was sitting in an intimate OpenAI builder town hall with about 50 people when I raised my hand: How can AI be used to solve long-standing economic gaps, like the gender wage gap?<\/p>\n
His answer was immediate. AI, he said, should be \u201can equalizing force in society.\u201d<\/p>\n
I agree. But right now, it isn\u2019t. Right now, AI is making the gender gap worse.<\/p>\n
The numbers are stark. A study published in the Proceedings of the National Academy of Sciences<\/a> found that women are 16 percentage points less likely than men to use generative AI tools for work. According to Deloitte<\/a>, only 28% of women report using AI regularly, compared to 45% of men. Mobile download data<\/a> shows women comprise just 27% of ChatGPT app users.<\/p>\n The gap is widest among the youngest workers. Among Gen Z<\/a>, 71% of men use generative AI weekly compared to 59% of women. These are the workers who will carry AI skills, or the lack of them, throughout 40-year careers. The compounding effect is staggering.<\/p>\n Meanwhile, the gender wage gap costs American women $1.6 trillion annually<\/a>. AI could help close it. Instead, we are watching the technology meant to democratize capability become another mechanism of inequality.<\/p>\n The problem is not that women cannot use AI. The problem is that the AI tools flooding the market were built by male-dominated teams for general-purpose use. They do not address the specific challenges women face in career advancement: negotiating salaries, building professional visibility, navigating workplace dynamics that research consistently shows disadvantage women.<\/p>\n Consider negotiation. A landmark study by Linda Babcock<\/a> found that only 7% of female MBA graduates negotiated their first salary, compared to 57% of men. That single decision at career entry can result in over $500,000 in lost earnings by age 60<\/a>. An AI tool purpose-built for women\u2019s salary negotiation, one that provides data-driven benchmarks, personalized scripts, and low-stakes practice environments, could systematically close this gap.<\/p>\n Or consider visibility. Research from McKinsey and LeanIn.Org<\/a> shows women are nearly twice as likely as men to be mistaken for someone more junior. They are 37% more likely to have colleagues take credit for their ideas. AI tools designed for personal brand building, content creation, thought leadership positioning, and network expansion could help women overcome structural barriers that have proven resistant to decades of diversity initiatives.<\/p>\n