The Hidden Bias in AI: Why People of Color Are Being Left Behind
AI is often presented as a technology built for everyone, but many AI tools still struggle with cultural nuance and representation. In this post, I explore the Western and Eurocentric biases I’ve encountered when using AI, particularly as a Black person, and why communities of colour need to play a more active role in building AI systems that reflect their experiences, cultures, and needs.
Temidayo Adefioye
I’ve had the opportunity to use various AI tools on multiple occasions, and I’ve noticed something consistent across the board. Tools like ChatGPT, among other powerful artificial intelligence technologies, struggle with cultural nuance.
As a Black person, this is particularly frustrating. It’s clear to me that people of colour are not properly represented in the algorithms that are often promoted as the future of technology. The current iteration of ChatGPT, for example, offers answers that are too generalized for specific questions tailored to certain communities, showing a training bias that appears Eurocentric and Western.
This isn’t an isolated issue — most AI models aren’t built with people of colour in mind. When I use ChatGPT, I don’t expect it to naturally respond in a way that resonates with Black people. To achieve a tone that feels authentic, I have to craft exceptionally detailed prompts, which is far from efficient.
Take, for instance, a simple query about important contemporary music artists. The model will likely mention Taylor Swift or Ed Sheeran, but it won’t bring up artists from India, China, Africa, or even African American artists. This bias highlights a focus on the European trajectory of history and culture, neglecting the rich diversity that exists globally.
However, I’ve noticed some startup companies addressing this problem by developing similar products that ensure people of colour are properly represented. This is a step in the right direction. As an AI company, it’s acceptable to specify that your product is designed for a specific region and to state this on your website. What isn’t acceptable is for any company to claim they are solving a global problem while failing to ensure that people of colour benefit equally from the tool.
Here’s my take: Africa and other communities of colour need to step up and get involved. We must start solving our problems because no one else will do it for us. We can’t rely on others to fix our issues. It’s our responsibility to equip our people with the knowledge and infrastructure required to tackle any challenges that arise in the future.
The path forward involves a concerted effort from within our communities to create AI tools and technologies that reflect our experiences, cultures, and needs. Only by taking charge of our technological future can we ensure that AI truly serves every person of colour.
Temidayo Adefioye
I’m a talent and workforce strategist focused on helping companies find exceptional people beyond the traditional CV. My journey into talent started by building teams, making hiring mistakes, and experiencing firsthand how difficult it can be to find people who can actually execute. That experience led to the creation of Talenvo, a platform connecting educators, early talent, and companies across Africa’s technology ecosystem. Through Talenvo, I’ve supported hundreds of companies across product, data, content, AI, design, and engineering by helping them build stronger talent pipelines and make better hiring decisions. Beyond Talenvo, I’m the author of three internationally published books and have trained more than 31,000 learners through LinkedIn Learning. I also write weekly on Substack about work, people, leadership, and the realities of building teams. I’m passionate about helping companies hire with less guesswork and helping emerging talent get opportunities to prove what they can actually do.
Comments
No comments yet. Be the first to share your thoughts!