AI and Identity

Lack of diversity in AI development directly shapes the technologies created: The paper argues that understanding "AI identity"—which encompasses who creates AI, what gets built, and who is affected—reveals how homogeneous creator communities produce biased systems. This suggests that increasing diversity among AI developers is not just an equity issue but a technical necessity for building fair AI.

The "AI divide" is widening, creating unequal impacts across society: Beyond the well-documented lack of diversity within the AI field itself, the paper identifies a growing gap in who benefits from AI versus who bears its risks. This divide means that marginalized communities may face AI's harms while having little say in its development.

A three-dimensional framework is needed for responsible AI: The paper proposes addressing AI bias and ethics through three interconnected areas—Creators (who builds AI), Creations (what AI systems are built), and Consequences (who is affected and how). This suggests that fixing AI responsibility requires simultaneous action across all three dimensions rather than isolated fixes.

Identity is central to understanding AI ethics: By framing the problem through identity—examining the associations between developers, technologies, and social impacts—the paper suggests that ethical AI requires understanding not just technical fairness metrics, but the human and social dimensions of who is involved and affected.