Women in AI Leadership: Fewer Than 14% of the Seats Are Ours. Here Is What to Do From Yours.

Erica Anderson Rooney is an ex-CPO turned executive coach and AI educator on a mission to get more women into positions of power and keep them there. She is the author of The AI Gap (2026), host of the Glass Ceilings & Sticky Floors and AI Voice or Victim podcasts, and NC Chapter Lead for Women Applying AI. Her SNAP and AI Continuum frameworks are taught to thousands of women through CHIEF, WICT, and Fortune 500 leadership programs. She is the sought-after speaker at the intersection of women, AI, and leadership.

Somewhere in your company this month, a small group of people sat down and decided how AI will be used on your team. Which tool gets added to the tech stack. What it is allowed to touch. Who checks its work. I would bet you were left off the invite, and I would bet most of the people who got one were men.

That is the state of women in AI leadership right now, and the United Nations just quantified it.

Let's get real for a second. You are a smart and capable woman with a big job, and you probably don’t have a lot of time to spend worrying about all the AI mess in your organization. You get the email announcing the new tool. You get the training link (maybe). You get the policy after it’s been written and pushed out.

So you do what capable women always do. You adapt. You figure out the tool on your own time, you make it work for your team, and you keep the wheels on.

Meanwhile, the decisions that shape your work, your team's jobs and your next promotion are being made a few floors up by people who have never done your job and they have no idea what it entails.

I spent more than 15 years in HR, all the way up to Chief People Officer. That work taught me one thing about every big change a company makes: the people in the room when a system gets designed are the people the system works best for. Everyone else gets the finished product and a deadline.

I will never forget when a parent company I worked for shoved a version of Workday down our throats as the new HRIS system. I loved what Workday could do, but of course the parent company did not purchase the version of the tech that would work for my subsidiary. So off to India I flew, for weeks at a time, trying to figure out how to make a square peg fit into a round hole. It was miserable. And it was not going to be the “employee experience” anyone had hoped for.

You're not crazy for feeling like AI is happening TO you. For most women, it is.

What the UN Numbers Say About Who Builds AI

On September 22, 2026, UN Women released new figures during the UN General Assembly's High-Level Week in New York. They come from its Gender Snapshot 2026 report, and they are worth reading slowly.

Women make up around 22% of AI professionals globally. Women hold fewer than 14% of senior leadership positions in AI. UN Women put it the other way around too: men hold nearly 8 in 10 AI jobs and almost 9 in 10 of the senior leadership positions shaping the technology.

The same release says women are around 20% less likely than men to use generative AI tools. (Generative AI means tools like ChatGPT, Claude and Copilot that create text, images or answers when you type a request.)

Now add the other half of the picture. In March 2026, the International Labour Organization published a research brief on generative AI and women's jobs. It found that around 29% of female-dominated occupations are exposed to generative AI, compared with 16% of male-dominated occupations. Exposure means how much of a job's daily tasks these tools could do or change.

It gets sharper at the top of the risk scale. The ILO found 16% of female-dominated occupations fall into the highest exposure categories, compared with 3% of male-dominated ones. Women were more exposed than men in 88% of the countries it analyzed.

Put those two reports side by side. The women with the most at stake in how AI gets used at work have the smallest say in how it gets built.

Sarah Hendriks, Director of Policy, Programme and Intergovernmental Support at UN Women, said it in one sentence: "The future of AI cannot be built by half the world's talent."

One day later, on September 23, 2026, the UN Security Council held a meeting on AI risks. Reuters named four AI leaders who spoke: Sam Altman of OpenAI, Dario Amodei of Anthropic, Clément Delangue of Hugging Face and the researcher Yoshua Bengio. All four are men.

Sam Altman told the council, "If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone." I agree with him. I would add that those decisions cannot be made by one half of the population either.

Why Are There So Few Women in AI Leadership, and Why Should You Care?

Here is my read on the why, as a former Chief People Officer.

Part of it is plain math. Senior AI leaders get picked from the pool of AI professionals, and that pool is about 22% women. By the time you reach the senior seats, the UN Women figure drops below 14%. Women thin out at every step up, a pattern anyone who has worked in HR will recognize.

(I talked through the pipeline side of this with Anna Radulovski, founder and CEO of Coding Girls, on Glass Ceilings & Sticky Floors. The episode is called Shattering the Silicon Ceiling, and it is a good listen if you want to go deeper.)

The other part is how companies staff the decision. Many organizations treat AI as a technical project, so they fill the AI committee with technical titles. The head of HR, the head of customer care, the woman who runs the team whose work is about to change? She hears about it at rollout.

I think that is a mistake, and I think it is one you can help fix.

Here is why you should care, even if you never plan to work in tech. "Building and governing AI" comes down to a series of ordinary human decisions.

Building means deciding what problem a tool should solve, what information it learns from, how it gets tested, and who it gets tested on. Governing means setting the rules: where the tool can be used, what it is allowed to decide without a person checking, who is accountable when it gets something wrong, and how anyone finds out that it did.

Every one of those is a judgment call. And judgment calls reflect the lives of the people making them.

History is written in the training data. AI systems learn from records of what already happened, and those records carry every old pattern about who got hired, who got promoted and whose work counted. Someone has to be in the room to ask, "Whose history did this learn from, and who did we test it on?" Hendriks made the same point in the UN Women release: when women are left out of AI's design, governance and regulation, today's inequalities risk getting embedded in the systems of the future.

This is already showing up in real harm. A UN Women study published in 2026, Tipping Point, found that almost one in four incidents of online violence against women human rights defenders, activists and journalists involved AI.

And here is the part that gives me hope. Senior women are already stepping into these decisions. In a 2026 Harris Poll survey for CHIEF, 80% of the 1,005 senior women leaders surveyed said they are active strategic players in their organization's AI efforts. And 85% had taken action inside their organizations, including establishing AI governance guidelines.

Read that again. Those are women at VP level and above, and they are helping write the rules.

In The AI Gap, I walk through Four Gaps that stand between women and power in the age of AI: Awareness, Authority, Compounding and Leadership. This news sits squarely in the fourth one, the Leadership Gap, which is all about women shaping AI while it is still being shaped. So here are five moves you can make from the seat you are in today.

Move 1: Find the Room Where AI Decisions Get Made at Your Company

You cannot join a room you cannot find. So this week, go find it.

Ask your boss, your IT partner or your HR partner three things. Who decides which AI tools we buy? Who wrote our AI policy, or who is writing it now? Is there a committee, a working group or a single person who signs off?

Then look at the list of names. Count the women. Count the people who run a team that does the daily work the tool will touch.

I have a hunch about what you will find, and so does the UN.

Here is what this looks like in practice. Say you run a 40-person customer care team and you learn that a pilot of an AI assistant for your agents was scoped by IT and finance. You send one email: "My team is the one using this every day. I would like to be part of the pilot group so the testing reflects how the work gets done." That email takes four minutes, and it is very hard to say no to.

Move 2: Ask the Four Governance Questions (No Coding Required)

Governing AI takes good questions and the nerve to ask them in a meeting. A technical background is optional.

Here are the four I want in your back pocket.

One: What did this tool learn from, and who was missing from that information? Two: Who did we test it on before we rolled it out? Three: What is it allowed to decide without a person checking? Four: When it gets something wrong, who is accountable, and how will we know?

Think about how these land in a real meeting. If your company is looking at an AI tool that helps screen resumes or summarize performance reviews, question two alone can change the outcome. Did anyone check how it handles a resume with a two-year caregiving gap? Did anyone compare its summaries for men and women doing the same job?

If the answer is silence, you just surfaced a risk nobody else in the room was looking for. And you did it with a question.

Move 3: Use the Tools Enough to Have an Opinion

Remember that women are around 20% less likely than men to use generative AI tools. That gap matters for your own productivity. It matters even more for your credibility in the decision room.

It is hard to argue about how a tool should be governed if you have never watched it get something wrong. And it WILL get things wrong. Seeing that with your own eyes gives you a point of view worth hearing.

So set a small, specific goal. Use one AI tool for 15 minutes a day for the next two weeks, on real work. Draft the tough email. Summarize the 30-page report. Ask it to poke holes in your budget proposal.

Keep a running note of two things: where it saved you time, and where it made a mistake you had to catch. That note is your credential. When you walk into the AI conversation with five specific examples from your own function, you are the most useful person there.

This is what I mean when I talk about knowing how to talk to the machine. It is a leadership skill, and it is learnable in the pockets of a normal week.

Move 4: Raise Your Hand for the AI Seat, and Bring a Woman With You

Once you know where the room is, ask to be in it. Out loud. In writing.

Try this: "I would like to join the AI working group as the voice of [your function]. My team's work is directly affected, and I can bring real use cases and real risks." Short, specific, and built around what you add.

If there is no working group yet, even better. Offer to start one. The person who convenes the conversation usually ends up shaping it, and that is authority you can point to at review time and when you negotiate your next raise.

Then do the thing that changes the math. When you get the seat, name another woman for the next one. Recommend the operations director who knows where every process breaks. Recommend the HR business partner who hears what employees are afraid to tell their managers.

One woman on an AI committee can be outvoted. Three can carry a decision.

This is also why I care so much about women doing this together, and why I built HER Collective. There is a post over there on seven ways to bridge the AI gap.

Move 5: If You Lead a Team, Protect the Most Exposed Jobs First

If you manage people, you are already governing AI, with or without the title. You decide how your team uses it, who gets trained and whose role changes first.

Go back to the ILO numbers. Female-dominated occupations are almost twice as likely to be exposed to generative AI as male-dominated ones. Think about the administrative, coordination and support roles on your team, and look at who holds them.

Those are the women who should get AI training FIRST. Give them the tools, the time to learn and a hand in redesigning their own work. A coordinator who learns to run the AI tool that handles scheduling is building her next job. A coordinator who watches it get rolled out around her is waiting for a meeting invite from HR.

Janine Berg, a senior economist at the ILO and co-author of the brief, said the impact of generative AI on women's jobs "is not predetermined." I love that line, because it puts the outcome back in the hands of the people making decisions. In your corner of the company, that person is you.

The Bigger Picture: AI Is a Power Shift

AI is a power shift. Big technology changes move money, authority and opportunity from one group of people to another, and the people who set the rules early tend to keep the advantage.

We are in the early part right now. The policies are in draft. The committees are being formed. The norms for what AI can and cannot decide about a person's career are being written this year, in rooms where men hold almost 9 in 10 of the senior seats.

I am telling you this because the rules are still in draft, and a draft can be changed.

The future of work will not wait for women to feel ready. But here is the good news: readiness, in this case, is a handful of questions, a daily habit and one email asking for a seat. You have done much harder things before 9 a.m.

What I Want You to Take From This

The UN numbers are real, and they are bad. Women are about 22% of AI professionals and fewer than 14% of senior AI leaders, while women's jobs carry almost twice the exposure.

You can change who governs AI in the place that matters most to your career, which is your own company. Find the room. Ask the four questions. Use the tools enough to have an opinion. Ask for the seat and bring a woman with you. Protect the most exposed jobs on your team first.

Every one of those moves builds your authority, and authority is what gets paid.

If you want the full playbook, that is why I wrote The AI Gap. It walks you through all Four Gaps, including the Leadership Gap we talked about today, with the frameworks and prompts to close each one. You can get your copy here.

I am cheering you on, all the way into that room.

Erica

Frequently Asked Questions

Why are there so few women in AI leadership?

UN Women's Gender Snapshot 2026 found that women are around 22% of AI professionals globally and hold fewer than 14% of senior AI leadership positions. Senior leaders are drawn from a pool that is already mostly men, and women thin out further at each step up. Inside companies, AI decisions are often staffed with technical titles, which leaves out many of the women whose teams are most affected.

Why does it matter who builds and governs AI?

The people who build and govern AI decide what problems it solves, what information it learns from, who it gets tested on and what it can decide without a person checking. Those are human judgment calls, and they reflect the lives of the people in the room. When women are missing, the needs and risks that affect women are less likely to be raised before a tool goes live.

How can I influence AI decisions at work if I am not in tech?

Start by finding out who decides which AI tools your company buys and who writes the AI policy. Ask to join that group as the voice of your function, and bring real examples from your own use of the tools. In The AI Gap, I call this closing the Leadership Gap: shaping how AI is used at work while the rules are still being written.

Erica Anderson Rooney is an ex-CPO turned executive coach and AI educator on a mission to get more women into positions of power and keep them there. She is the author of The AI Gap (2026), host of the Glass Ceilings & Sticky Floors and AI Voice or Victim podcasts, and NC Chapter Lead for Women Applying AI. Her SNAP and AI Continuum frameworks are taught to thousands of women through CHIEF, WICT, and Fortune 500 leadership programs. She is the sought-after speaker at the intersection of women, AI, and leadership.