AI Can't Be What AI Can't See
“AI Can’t Be What AI Can’t See” coined by Eli Potter, Author and Technology Executive Advisor, in her book Role Modelship is a North Star for the future of humanity.
AI is a giant mirror. And right now, it’s reflecting a fractured image.
AI learns from us our histories, our habits, our choices. Feed it a world where certain voices are missing, certain faces are absent, certain stories are untold and it won’t just fail to represent them. It will erase them. At scale. At speed. Permanently baked into the systems that will shape hiring, healthcare, education, and opportunity for generations.
Just as a child raised without diverse role models cannot imagine possibilities beyond their narrow world, AI trained on incomplete, biased, or homogeneous data cannot model, reflect, or empower the full range of human potential.
This isn’t a technical problem. It’s a moral one.
We have one window to get this right to ensure that the most powerful technology ever built reflects the best of humanity, not just the loudest, most documented, most privileged slice of it.
The data we feed AI today is the world it will build tomorrow.
For decades, role modelship was filed away under inspiration. Mentorship. Culture. The intangible stuff. Nice to have. Easy to deprioritize.
That era is over.
Every human behavior that gets documented, published, filmed, cited, or shared becomes training data. It becomes the raw material that AI systems consume, learn from, and replicate. Who we hold up as leaders, whose stories we tell, whose expertise we amplify, whose faces appear in positions of power and possibility — all of it flows directly into the models being built right now.
This means role modelship is no longer just about inspiring the next generation. It is about programming it. Role models are identity engineers.
When a young woman in medicine is never featured in the data, AI doesn’t just fail to see her. It learns she doesn’t belong there. When women are underrepresented in tech literature, AI doesn’t just reflect that gap. It calcifies it. When leadership is consistently modeled by a narrow demographic, AI doesn’t just notice the pattern. It enforces it.
The humans we choose to model excellence, the stories we choose to elevate, the representation we choose to normalize — these are no longer cultural decisions. They are architectural ones. They are inputs into a system that will make consequential decisions about human lives at a scale no individual bias ever could.
Role modelship was always important. Now it is infrastructure.
Who we celebrate shapes what AI believes is possible. And what AI believes is possible shapes the world it builds for humans.