Purposeful AI

Abstract

The potential of AI to improve people's lives is unlimited. However, designing AI-infused systems to realize this potential for all is challenging, and it is particularly challenging within under-served minority communities. Revisiting algorithms and data sources to eliminate biases in AI is an important step toward expanding its benefits. However, this leads to the realization that some applications, although unbiased, may simply not be what marginalized communities need. Marginalized communities live in a unique socioeconomic and cultural context and have difficulties taking advantage of traditional AI-powered systems due to language, digital literacy, or other barriers. Additionally, the systems that currently exist usually fail to address the unique challenges in these communities. Therefore, designers and researchers in AI must understand that a fundamental dimension of fairness and ethics is to empower and lift the lives of under-served populations through purposeful human-AI research. This research not only recognizes the unique intersectionality and needs present in each community, but also treats them as the primary audience by overcoming traditional barriers between researchers and underserved communities. This SIG proposal aims to initiate a multidisciplinary discussion around the design of AI systems that are purposefully targeted to marginalized populations. Within this discussion, our objective is to better understand how to conduct research to support and serve these communities - particularly if we are not members of such communities; and challenge the effectiveness of techniques like user-centered design in the context of AI-infused systems when the designer is not a member of the user community. In addition, we will delve into ways of translating research into practical applications to create a positive impact within these communities and narrow the disparities that exist in the optimal utilization of AI to improve lives for different communities.