Key Takeaways: Generative AI and Black Communities

• Scholarly research on AI bias focuses too narrowly on technical fixes: The paper's review of 91 empirical studies found that academic research concentrates heavily on detecting technical bias in datasets and proposing technical reforms, rather than examining how generative AI affects Black communities' actual lived experiences, cultural practices, or structural conditions.

• A dangerous gap exists between public and academic understanding: Public discourse recognizes that AI harms stem from "historical and systemic forces," but scholarly research stops its analysis at the dataset level. This misalignment means the solutions being developed in academia don't address the root causes identified in public conversation.

• Black knowledge and perspectives are being excluded from AI development: The paper argues that both public discourse and scholarly research fail to center Black epistemic agency—meaning Black communities' own ways of knowing and understanding are not being incorporated into how AI systems are designed, evaluated, or deployed.

• Current AI ethics approaches are insufficient: The authors demonstrate that standard technical evaluation methods foreclose important questions about cultural impact and structural inequality, suggesting that new methodologies (like frame analysis) are needed to surface what purely technical approaches miss.

• Reducing Black identity to "measurable variables" obscures real harms: By treating Blackness as data points to be detected and corrected, research misses how genAI systems interact with Black communities' actual social, economic, and creative contexts in education, employment, healthcare, and other sectors.