Invited Symposium II: Algorithms, Bias, and Fairness [presented in person and live streamed]

Tools from the world of artificial intelligence (AI), including machine learning algorithms (MI), have become ubiquitous. They have also begun to move beyond computer science, where they originated, into other fields, including social work. This session will address four questions about the deployment of machine learning methods/techniques: 1) what are some of the main methods/techniques which are used, 2) what sorts of questions are they used to address 3) how, specifically, has the work of the two panelists drawn upon these methods/techniques, 4) what are some of the ethical issues raised by the use of these methods/techniques?

The fourth question is especially salient, given social work’s commitment to social justice as well as the documented cases of algorithmic bias/unfairness when these methods/techniques have been deployed outside of social work. For example, ProPublica reported that COMPAS, an algorithm deployed to predict the likelihood of reoffending tends to overestimate reoffending rates for Blacks and underestimates them for White. Another example is facial recognition software which appears less likely to accurately classify people with dark skin. Given such examples, the session will focus on what social workers who use such techniques might be able to prevent these problems from occurring in our field.