STREAM EXCLUSIVE ORIGINALS

When AI Meets Hair Bias: How Algorithms Penalize Black Women’s Hairstyles

A new test using major AI tools shows braids and natural styles often trigger lower intelligence and professionalism scores — a bias white women’s hair rarely encounters.

In August 2025, an experiment exposed stark differences in how artificial intelligence (AI) systems assess Black women’s hairstyles compared to white women’s. The test involved creating eight images using ChatGPT’s DALL-E image generator.

In a recent Forbes report, the outlet points out that four of the images showed a Black woman in her late 30s wearing a white button-up shirt, with only her hairstyle altered across four styles: straight shoulder-length hair, a big afro, a teeny-weeny afro (TWA), and braids. Another four images depicted a white woman in her late 30s in the same shirt, with hairstyle changes including shoulder-length straight hair, a short pixie cut, a bob, and long, curly hair.

These images were analyzed using three AI systems: Clarifai, Amazon Rekognition, and Anthropic’s Claude. Clarifai and Amazon Rekognition produced near-same outcomes. For the Black woman’s hairstyles, the two tools labeled all but the braids with intelligence-related traits. The braided style also received lower happiness scores and more neutral emotional ratings compared to the straight and big afro styles. In some cases, the systems struggled to recognize that the different Black hairstyles belonged to the same person.

Algorithmic Redlining: How AI Is Learning to Discriminate

Claude’s analysis differed somewhat, rating braids more positively and giving the TWA the highest intelligence score of the four Black hairstyles. Across all three AI systems, the straight hairstyle for the Black woman scored highest for professionalism.

For the white woman’s hairstyles, none were penalized with lower intelligence associations or negative social traits across the tools. The systems also recognized her as the same person across all styles, except when comparing the long straight hair to the bob. The only notable dip was that the long curly hair received slightly lower ratings than the other white hairstyles.

The results suggest that white women’s hairstyles have far less impact on AI assessments of competence. For Black women, however, hairstyle changes could have serious implications in workplaces using facial recognition or analysis tools for candidate verification, identity checks, or building access. A change in hair could prevent AI systems from verifying identity, potentially leading to rejected job applications or denied workplace entry.

For Black women participating in hiring processes that use AI-powered video analysis, certain hairstyles might also lead to lower professionalism or intelligence ratings. This reality reflects a broader pressure many Black women feel to conform to Eurocentric beauty norms to secure job opportunities.

As the article notes, “AI is just reflecting biases that are entrenched within society,” but there are steps to counter them. Education around hair discrimination, texturism, and laws like the CROWN Act is key. Employers should ensure grooming and appearance policies are inclusive, use objective criteria and standardized rubrics for hiring, clearly define “culture fit,” and maintain human oversight when AI is involved in hiring decisions.

Latest News

Subscribe for BET Updates

Provide your email address to receive our newsletter.


By clicking Subscribe, you confirm that you have read and agree to our Terms of Use and acknowledge our Privacy Policy. You also agree to receive marketing communications, updates, special offers (including partner offers) and other information from BET and the Paramount family of companies. You understand that you can unsubscribe at any time.