Towards a Feminist Metaethics of AI

AI ethics guidelines alone are insufficient: The paper argues that despite the proliferation of AI ethics codes and principles, these outputs have proven inadequate, as demonstrated by "recent problematic incidents about AI ethics/ethicists," suggesting that having ethical frameworks on paper doesn't guarantee responsible AI development in practice.

AI ethicists need to examine their own field: The author contends that before AI ethicists evaluate other professions and industries, they should "critically evaluate their own" practices and processes more systematically—a self-reflection that is currently lacking in the literature.

Feminist metaethics expands beyond abstract principles: Rather than just asking "what is ethics," a feminist metaethics of AI asks "what our engagement with it should be like," shifting focus from theoretical debates to how ethics is actually practiced and experienced by real people affected by AI systems.

Real-world impact and power dynamics matter: The proposed research agenda emphasizes examining the actual effects of AI ethics work on people's lives and how AI affects power relations in society, rather than focusing solely on competing moral theories and values.

Context, emotions, and lived experience are essential: The paper advocates for methods that pay attention to context, emotions, and personal narratives when evaluating AI ethics, moving beyond abstract philosophical analysis to understand how AI decisions affect different communities and individuals.