Explaining the ghosts: Feminist intersectional XAI and cartography as methods to account for invisible labour

AI systems depend on massive amounts of hidden human labor: The paper emphasizes that contemporary AI automation relies on substantial behind-the-scenes work—such as data labeling and system maintenance—that users typically never see or think about, yet this labor is often underpaid and undervalued.

Invisible labor should be made visible to users: The authors argue that explainable AI (XAI) design should be used to sensitize users to the role of human labor in AI systems, suggesting this transparency is an important part of responsible AI development.

Feminist intersectional approaches offer new perspectives on AI accountability: The paper proposes using feminist intersectional XAI methods to better account for dimensions of AI systems that are typically overlooked, particularly those related to labor and power dynamics.

Cartography can map the full AI system: The authors introduce cartography—a method from feminist intersectional research—as a tool to create a more complete, systemic view of how AI systems actually work, including the human labor that makes them function.

Current XAI methods may be incomplete: By highlighting invisible labor as a critical gap, the paper suggests that standard explainability approaches may not be telling the full story about how AI systems operate or who contributes to their functioning.