-
Expand entry
-
Added by: Clotilde Torregrossa, Contributed by: Marilou NieddaAbstract:
Artificial intelligence (AI) systems are increasingly adopted to make decisions in domains such as business, education, health care, and criminal justice. However, such algorithmic decision systems can have prevalent biases against marginalized social groups and undermine social justice. Explainable artificial intelligence (XAI) is a recent development aiming to make an AI system’s decision processes less opaque and to expose its problematic biases. This paper argues against technical XAI, according to which the detection and interpretation of algorithmic bias can be handled more or less independently by technical experts who specialize in XAI methods. Drawing on resources from feminist epistemology, we show why technical XAI is mistaken. Specifically, we demonstrate that the proper detection of algorithmic bias requires relevant interpretive resources, which can only be made available, in practice, by actively involving a diverse group of stakeholders. Finally, we suggest how feminist theories can help shape integrated XAI: an inclusive social-epistemic process that facilitates the amelioration of algorithmic bias.
Comment: It is an interesting text for students who work specifically with the concept of XAI (Explainability) in phil of AI. It offers an alternate perspective of this super technical/analytical concept. Can be used by students in computer science (in an ethics class, perhaps).
-
Expand entry
-
Added by: Tomasz Zyglewicz, Shannon Brick, Michael GreerIntroduction: Let’s set the stage. In 2016, ProPublica released a ground-breaking investigation called Machine Bias. You’ve probably heard of it. They examined a criminal risk prediction tool that’s used across the country. These are tools that claim to predict the likelihood that a defendant will reoffend if released, and they are used to inform bail and parole decisions.Comment (from this Blueprint): This is a written transcript of the James Baldwin lecture, delivered by the computer scientist Arvind Narayanan, at Princeton in 2022. Narayanan's prior research has examined algorithmic bias and standards of fairness with respect to algorithmic decision making. Here, he engages critically with his own discipline, suggesting that there are serious limits to the sorts of quantitative methods that computer scientists recruit to investigate the potential biases in their own tools. Narayanan acknowledges that in voicing this critique, he is echoing claims by feminist researchers from fields beyond computer science. However, his own arguments, centered as they are on the details of the quantitative methods he is at home with, home in on exactly why these prior criticisms hold up in a way that seeks to speak more persuasively to Narayanan's own peers in computer science and other quantitative fields.