A Systems Thinking Approach to Algorithmic Fairness
It addresses fairness in AI for policymakers, offering a sociotechnical framework to align policies with political agendas and democratic values, but appears incremental as it combines existing techniques.
The paper tackles algorithmic fairness by modeling bias in data generation using systems thinking and causal graphs, linking AI systems to politics and law to inform policy design.
Systems thinking provides us with a way to model the algorithmic fairness problem by allowing us to encode prior knowledge and assumptions about where we believe bias might exist in the data generating process. We can then encode these beliefs as a series of causal graphs, enabling us to link AI/ML systems to politics and the law. This allows us to combine techniques from machine learning, causal inference, and system dynamics in order to capture different emergent aspects of the fairness problem. We can use systems thinking to help policymakers on both sides of the political aisle to understand the complex trade-offs that exist from different types of fairness policies, providing a sociotechnical foundation for designing AI policy that is aligned to their political agendas and with society's shared democratic values.