The Governance Gap Hiding in the Gender Gap
The most governance-ready employees in your organization are probably the ones you're trying to "fix."
Women are adopting generative AI at roughly 20 percent lower rates than men. That's the finding from a Harvard Business School working paper by Nicholas G. Otis, Solène Delecourt, Katelyn Cranney, and Rembrand Koning, drawing on 18 studies and more than 140,000 participants worldwide. The typical response I hear from leaders: We need to get women to use AI.
That's not wrong. But if that's all you see, you're missing the more valuable signal.
It's Not Just an Access Problem
Otis and colleagues tested this directly. In a field experiment with more than 17,000 adults in Kenya, participants were given the opportunity and guidance to learn how to access ChatGPT, yet women were still about 13 percent less likely than men to engage with the tool. The research points instead to ethical concern and reputational risk — women appear more worried about whether relying on AI-generated work will be perceived as cheating or cutting corners. And research from Stanford's Shelley Correll shows why that worry is rational: women are often held to stricter standards of proof before they are judged equally competent. The cost of anything that calls your expertise into question is simply higher if you're a woman.
But there's something else going on too. At a WomenxAI panel earlier this year and in communities like Old Girls Club, I've watched women raise a different set of questions entirely:
Who is accountable when the output is wrong?
What happens to junior employees whose work this displaces?
How do we know the training data isn't encoding biases we've spent decades trying to dismantle?
Those aren't symptoms of technophobia. They're symptoms of thinking two moves ahead.
Meanwhile, Governance Is the Road to ROI
The Larridin 2026 State of Enterprise AI report surveys 364 enterprise leaders across 16 industries. It finds that while 69 percent of organizations report having AI policies, 37 percent admit their governance is inconsistent, 45.6 percent don't know their workforce adoption rate, and 37.1 percent say risk visibility is unknown. And organizations with formalized AI risk and compliance policies are 2.2 times more likely to demonstrate ROI than those without (84.5 percent versus 37.9 percent).
As Larridin put it, the era of "move fast and break things" is over. But their version of governance is the floor — tool inventory, policy enforcement, spend visibility. Organizations also need the harder layer: ethical boundaries, accountability structures, and an honest reckoning with downstream effects on people. That's the governance layer most organizations haven't built yet.
The Irony: You Already Have Governance Talent
A significant portion of the workforce is already asking governance-level questions. They're disproportionately women. And the dominant response is to treat their questioning as a participation problem — more training, more encouragement to just start clicking.
The women who are hesitant aren't on the sideline because they lack capability or curiosity. Many are weighing costs and consequences that eager early adopters are not. Their orientation toward ethics, accountability, and community impact isn't a barrier to AI transformation. It's a capability most organizations are paying consultants to build from scratch. In other words, the same risk sensitivity that suppresses early adoption in some contexts is exactly what you need in the room when you're deciding how AI should be used at scale.
What to Do Differently
Stop framing the gender gap as purely an adoption problem. Bring ethically attuned employees — including non‑adopters — into governance design conversations. Their concerns are your risk register.
And it's worth noting: in one tech study Otis and colleagues review, women in senior technical roles were actually slightly more likely than men to use generative AI, while junior women still lagged behind junior men. That's a clue that when exposure, expectations, and psychological safety line up, the gender gap doesn't just shrink — it can flip. The gap isn't fixed. It's contextual. The organizational environment, not the individual, is where the intervention belongs.
The caution isn't the problem. The failure to listen to it is.
Further reading
Otis, N. G., Delecourt, S., Cranney, K., & Koning, R. (2024). Global Evidence on Gender Gaps and Generative AI(Working paper). Harvard Business School.
Correll, S. J. (2004). Constraints into preferences: Gender, status, and emerging career aspirations. American Sociological Review, 69(1), 93–113.
Larridin. (2026). State of Enterprise AI Report, Q1 2026.
