AI and the Future of Racial Justice

By Culturati Scholar, Professor Craig Watkins, MIT & University of Texas at Austin

Attribution: University of Texas at Austin.

Many cities and organizations are adopting technologies like artificial intelligence (AI) to guide how departments organize data, make decisions, and deliver services to citizens. The widespread adoption of these systems is propelled by the desire to automate decision-making, manage large scale databases, and reduce the likelihood of human bias in decisions related to the allocation of resources and services. However, even as these technologies augment human labor and lead to greater workplace efficiency a critical question emerges: to what extent do these systems practice bias and replicate inequities?

Because AI systems rank, profile, and sort, they have the power to shape allocative practices (Eubanks, 2018). In the context of local governments and organizations, this means, for example, that AI systems can determine who gets access to employment, care, or adequate housing. Without the proper mechanisms to evaluate and test these systems through an equity lens, they may replicate systemic inequalities.

This interdisciplinary research collaboration will identify, propose, and develop ways to systematically address the inclination for bias and disparate impacts that characterize AI systems. We will do this in the following ways:

• In consultation with the City of Austin’s Equity Office our team will develop and implement a pilot research collaboration with at least one department to identify and propose solutions to address disparate impacts in the adoption and deployment of technology.
• Through measuring and understanding racial disparities in treatment and service provision related to the use of AI in key areas of life such as public safety, transportation, and health. Furthermore, we will produce a set of recommendations and solutions that mitigate bias and racially disparate impacts in the application of data-based systems.
• Through the building of “good systems” and “action plans” that are informed by interdisciplinary approaches and designed to expand access to critical life-affirming services.

Finally, the core feature of our research activities centers racial justice and equity as the main locus of study and problem-solving interventions. Moreover, our project is designed to realize one of the key goals of the Good Systems Grand Challenge, which is to pursue scholarly inquiries that engage community stakeholders and strive for real world impact.