Response to "Impossibility of What?" Publication

To start off, first, I would like to say that the publication that I read is pretty interesting and worth reading. The author, Ben Green, touch upon on how the algorithmic fairness provides the appropriate conceptual and practical tools for enhancing social equality. He argue that the dominant, "formal" approach to algorithmic fairness is ill-equipped as a framework for pursuing equality, as its narrow frame of analysis generates restrictive approaches to reform. Green does the distinction between formal and substantive algorithmic fairness is exemplified by each approach's responses to the "impossibility of fairness." While the formal approach requires us to accept the "impossibility of fairness" as a harsh limit on efforts to enhance equality, the substantive approach allows us to escape the "impossibility of fairness" by suggesting reforms that are not subject to this false dilemma and that are better equipped to ameliorate conditions of social oppression.

Green begins his introduction with the real-life situation about how the U.S. is overlapping with crises of social and economic inequality. He chose Black American as an example for the inequality that they posses far fewer economic and political resources than their white counterparts and suffer disproportionately from the injustices of policing and mass incarceration. Another example is gender inequality where women are underrepresented in prominent jobs and leadership roles across society and face a significant pay gap relative to men. And as economic inequality rises, fewer people have access to the material resources necessary for a comfortable and dignified life.

So how is algorithm works in decisions socially and economically? In general, algorithms represent a novel approach to overcoming the cognitive limits and social limits of human decision-makers. Proponents describe how algorithms could "disparately benefit" historically disadvantaged groups, which are typically judged unfavorably by stereotype-prone human decision-makers. The general optimism toward algorithms by prominent and influential proponents has led to algorithms being used to improve the fairness of decision-making in contexts such as criminal sentencing, hiring, and social services. It also used to combat discriminatory decision-making which accumulates evidence that algorithms themselves can discriminate. The growing list of examples demonstrate algorithms making more favorable decisions for men than women and white individuals than Black individuals. Concerns about algorithmic bias have gained particular traction as data-driven algorithms became more pervasive in government, business, and daily life. These concerns have prompted the emerging field of "algorithmic fairness."

Throughout his analysis, he mentions the problems of "impossibility of fairness" by escaping it. Before escaping it, the formal approach has navigating the "impossibility of fairness" where it commits to sufficiency as the appropriate definition of fairness, aligning with notions of formal equality by emphasizing that fairness entails treating people the same based on their likelihood for future crimes. Despite its broad appeal, the formal equality response overlooks both the social context of why different people pose different crime risks and the morality of existing policy responses to high-risk individuals. On the substantive approach, this approach focusing on escaping the "impossibility of fairness" where it requires a substantive approach to algorithmic fairness that analyzes inequality and decision-making within the broader context of social relations of hierarchy. This approach covers the problems that exist in "impossibility of fairness" and given direct responses to them.

In conclusion, this publications explains the formal and substantive approach of "impossibility of fairness" pretty well and very broadly. They include examples that represent the "impossibility of fairness" such as gender inequality where women are underrepresented and men has high salary than women.

Reference:

Green: Impossibility of What:

Green - Impossibility of What.pdf

Download PDF • 297KB