A Critical Pragmatism Approach for AI Ethics: Lessons from Urban Planning Theory

Algorithmic fairness is a "wicked problem" that cannot be solved through technical solutions alone. The paper argues that AI ethics should recognize fairness challenges as inherently complex, value-laden, and intractable—similar to urban planning problems—rather than treating them as purely mathematical puzzles with definitive answers.

Urban planning theory offers practical lessons for AI ethics that technical approaches overlook. Critical pragmatism from planning emphasizes governance, resource allocation, stakeholder engagement, and navigating conflicts rooted in deep-seated differences—dimensions that purely technical fairness definitions miss but are essential for real-world implementation.

Reflective and deliberative approaches are necessary for addressing AI ethics in practice. Rather than imposing top-down fairness frameworks, the paper recommends that AI practitioners adopt critical pragmatism's reflective methods that consider what stakeholders actually do when facing conflicts and power imbalances, as demonstrated in case studies on mortgage lending, school choice, and feminicide data collection.

Stakeholder engagement and power dynamics must be central to algorithmic fairness work. The paper emphasizes that ethical AI development requires explicit attention to who has decision-making authority, whose values are represented, and how conflicts between different stakeholders' interests are resolved—not just technical metrics of fairness.