ai gender bias

Why AI Thought I Was a Man in a Hoodie

I asked AI to create an image of me using only my words. It came back with a man. This isn’t about technology gone wrong. It’s about what our systems quietly assume when no one tells them otherwise.

You can learn more than you might think when using AI.
Like misogyny, for instance.

After seeing a post on LinkedIn where a businesswoman asked AI to create an image of her based solely on her words — no descriptors, no hints — I got curious. The image she received? A tech bro.

So I tried it myself. I asked ChatGPT to create an image of me based only on what I’ve written and shared over time.

Surprise.
I also got a man. In a hoodie.

Not quite full-blown tech bro, but close enough to notice.

Never one to let things lie, I asked ChatGPT to explain how it came up with that result.

Here’s what it said my “character energy” looked like, based only on my words:

• Intelligent
• Competent
• Grounded and candid (no guru-speak)
• Protective and advocacy-oriented
• Tech-comfortable but human-first
• Witty, a little sarcastic, not mean about it
• “I’ve been around the block and I’m still curious” energy

None of that reads as inherently male to me. I know plenty of women who carry those traits. So why did AI automatically assume a man?

The short answer is that it isn’t AI’s fault.

It’s society’s.

AI is trained on data, and that data comes from the internet. When you zoom out far enough, the pattern becomes hard to ignore: competence, intelligence, authority, and credibility are still most often presented and reinforced as male.

As ChatGPT and I broke it down further, it explained that what we were dealing with was model bias. AI gender bias. A default created by training data where:

“approachable + competent + intelligent + modern”
quietly equals
“a young guy in a hoodie.”

It also clarified something important.

My actual writing voice does not read as:

• tech bro
• hustle-y
• growth-at-all-costs
• hype-driven
• jargon-forward
• “move fast and break things”

In fact, my writing consistently pushes back against that entire worldview.

So why the mismatch?

Because without guidance, AI doesn’t know how to step outside its defaults. It fills in the gaps using the patterns it’s been shown the most. And those patterns still lean heavily male.

Why am I talking about this?

It might seem like a strange thing to fixate on, but hang with me. There’s a larger point here.

We like to tell ourselves that because women are more visible in business today, we’ve “come a long way.” In reality, we haven’t come as far as we think. This small AI experiment makes that painfully clear.

Women in business still aren’t viewed the same as men. They aren’t paid the same. They don’t receive the same baseline respect. And when women do reach positions of authority, it’s often because they learned to play by male rules. They hardened their edges. They downplayed their femininity. They adapted to the system instead of the system adapting to them.

And even then, many still deal with constant mansplaining from male peers.

When an AI model automatically assumes that intelligence and competence must belong to a man, it’s telling us far more about our culture than it is about technology. It’s reflecting the content it was trained on. And much of that content is still deeply misogynistic, even when it doesn’t look overtly hostile.

It’s more than a man in a hoodie.

ai gender bias
My AI generated image based on words alone

This wasn’t about me being mistaken for a man. I’m not offended. I’m not even surprised. What’s worth paying attention to is how easily and invisibly that assumption happened.

AI didn’t read my words and decide women can’t be intelligent or competent. It simply mirrored what it’s seen over and over again. That intelligence, authority, and credibility are most often trusted when they come wrapped in a male package.

Bias doesn’t always show up loudly. Sometimes it shows up quietly, as a default. As an assumption no one thinks to question until something bumps up against it.

This experiment didn’t teach me anything new about myself. But it did offer a clear snapshot of the cultural water we’re all swimming in. One where women’s competence still has to be explained, qualified, or softened. One where neutrality still looks suspiciously male. And one where we’re quick to say “things are better now” without noticing how much invisible scaffolding is still holding the old assumptions in place.

If AI can help us see that more clearly, then maybe its most useful role right now isn’t just generating content, but exposing the patterns we’ve learned to stop noticing.

And once you see a pattern, you can’t really unsee it.

Follow for Biz Tips & Conversations That Matter!