I remember the first time I sat through an AI panel and counted speakers: nine men, one moderator. You felt the tilt in the room before the first slide. I’m telling you this because those small tilts become highways.
I’ve spent years watching technology conversations steer public policy and product design. You don’t need a PhD to see when a map has been drawn by one hand: the blind spots show up in who gets hired, what features launch, and whose risks are named. This matters because influence is a quieter form of power than code — it decides what questions get asked and which people get to answer them.
Newsrooms and labs quote men far more than women — that skews what we think AI is
On podcasts, in stories, and at conferences, male voices dominate the narrative. Harvard’s new survey of more than 5,000 Americans found media quotes and expert bylines tilt heavily male, far beyond the actual gender mix in AI roles.
When reporters, platforms like Google and OpenAI, and corporate press rooms repeatedly amplify men, the public hears a version of AI written by men for men. That becomes the shorthand investors, policymakers, and product teams use when they make bets on the technology.
Why are women underrepresented in AI?
The short answer is not a single failure but a stack of small barriers: hiring practices favoring certain networks, less access to workplace AI training, and pipeline leaks where women leave or never reach technical roles. Harvard’s report found women are less likely to be in AI development jobs and less likely to receive dedicated training at work — which compounds over time.
In product teams, men use AI more — which changes what gets built
At companies, men report higher AI usage both on the job and at home. That’s not neutral.
Usage shapes ergonomics, default assumptions, and priority features. If the people steering design are the primary users, the product fits them best. The result can be a product that performs brilliantly for one group and poorly for another — sometimes causing real harm, from misclassifying health data to generating abusive deepfakes aimed mostly at women and children.
Do men use AI more than women?
Yes, survey data shows men report higher day-to-day AI use and more workplace training. Women, in contrast, report more anxiety and lower trust, which reduces adoption and influence — a feedback loop that keeps women out of the decision seats.
In meetings you’ll hear certainty — but confidence doesn’t equal knowledge
I sat through focus groups where men claimed familiarity with AI names and concepts that didn’t exist. That was revealing.
The Harvard researchers measured exactly that: men and women answered AI knowledge questions incorrectly at similar rates, but men were likelier to assert false knowledge. When a panelist confidently cites a nonexistent tool, teams take that cue. Confidence can masquerade as expertise, and poorly sourced certainty can steer products and policies down the wrong path.
How does gender bias affect AI outcomes?
When one group defines problems and metrics, models encode those priorities. That can reinforce inequality — from hiring systems that screen out female-coded language to content filters that fail to catch the kinds of abuse women face. Media and product teams that echo only one perspective make these failures harder to detect.
Communities near data centers and clerical jobs feel the consequences first — the costs aren’t evenly distributed
Surveys from Morgan Stanley and others show women are more wary of nearby data centers and of AI’s local impacts. That concern is rooted in reality.
Generative AI has produced nonconsensual sexual deepfakes, primarily of women and children. At the same time, automations threaten administrative and clerical roles that skew female, creating a twin pressure of privacy harms and economic displacement. When infrastructure and policy are decided by people who feel optimistic about AI, those living with the downside often don’t get a say.
I want to be clear: more women at the decision table benefits everyone. Diverse input reduces the risk of collective blind spots — and it can surface problems before they scale. Think of it as a theater where only half the seats are visible; the performance might still go on, but you’ll miss entire sections of the audience unless you change the lights.
Investor rooms, boards, and product roadmaps are where futures are chosen — so who sits there matters
Laura Manley, co-author of the Harvard report, put it plainly: The question now is who gets to decide on the future of AI and how much will women be involved.
If men are more likely to claim knowledge they don’t have, and women are more likely to be cautious or excluded from training, the governance debates will reflect those imbalances. Names that get funding, leaders who set safety standards, and regulators who craft rules will all be influenced by this gendered mix of optimism and skepticism.
We can fix it in obvious ways: broaden hiring, fund training for underrepresented teams, and demand balanced expert sourcing in media. Tools and platforms — from OpenAI’s APIs to enterprise features at Microsoft and Meta — will adapt faster when diverse voices ask different questions. Companies like Instacart are already embedding AI into everyday services; the design choices they make today ripple into daily life.
I’ll tell you what I see as an editor and a skeptic: the loudest expert isn’t always the most informed, and a field that lets confidence trump accuracy will make expensive mistakes. Those mistakes cost money and trust — and sometimes safety. Will you let a single chorus define the future of AI, or will you insist the microphone be passed around?