Stevens Blogpost 3

 In her paper, “Are Algorithms Value-Free?”, Professor Johnson shows that all science, and consequently all computer algorithms, require assumptions or values to build knowledge. Scientists and algorithms use induction to draw conclusions from evidence. However, there is no correct way to link evidence to a certain conclusion. As Johnson says, “There are many possible bridges one might adopt to traverse the gap between evidence and theory, and there seem to be no a priori grounds for preferring some bridges over others.” (Johnson 5). Still, science has attempted to find some value-free standards to serve as objective, independent grounds for theory, such as accuracy, consistency, simplicity, and others.

Failing to recognize the relativism within scientific theory has resulted in some significant missteps. For example, Johnson discusses the case of Ambien in which clinical trials failed to consider metabolic differences between men and women, resulting in the FDA suggesting women take twice the correct dosage. In this instance, the scientists’ mistake was caused by privileging the value of simplicity over flexibility or ontological heterogeneity. Professor Johnson writes that the “scientists’ decision to posit the fewest kinds of entities in a context where members of a privileged class… …lead to theories that both legitimate and perpetuate the socio-political values on which they’re built.” (Johnson 9). This example reminds me of Cheryl Harris’s idea in her paper “Whiteness as Property” when the “settled expectations” of the dominant class are justified by science” In Harris’s case, the values of science in the Jim Crow era were reinforced by the background of whites to protect the racial hierarchy and whiteness as property.

 

In the second half of Johnson’s paper, she discusses the COMPAS algorithm that is used by judges in the US to access recidivism risk. While the program claimed to fair and objective, advocates showed that it was almost twice as likely to falsely label black defendants as future criminals than white defendants. Johnson shows that COMPAS falls prey to many of the same illusions about value-free science that expect a universal standard of fairness. There are many possible definitions of fairness, and reasonable methods of inference pull in opposite directions.

Johnson’s paper raises some interesting insights that I would not have considered, but in some ways it leaves me feeling a little bit hopeless. If both algorithms and human judges rely on subjective modes of induction, what basis do we have to prefer one over the other? Is there anyway to judge whether any particular reform is more “fair” than another?