Artificial Intelligence for Whom? Ethical and Cultural Considerations for Artificial Intelligence

• In Event: Artificial Intelligence in Education: Potentials and Challenges in Ethics, Data Privacy, and Pedagogy

Abstract

Designers of technology are all rooted in particular cultural, historical, and political knowledges and values. This is evident in calls to action such as ‘design or be designed’ (inspired by Norman, 1993) and “program or be programmed” (Rushkoff, 2010). As artificial intelligence (AI) has become more ubiquitous, these calls to action have become more critical in ensuring equity and justice in not only our present but also our futures. Leaders in the tech industry and education (e.g., Buolamwini & Gebru, 2018; Eubanks, 2018; Noble, 2018; O’Neil, 2016) have persistently spotlighted the harmful consequences of current forms of AI. Thus, it is evident that current forms of AI perpetuate systemic injustices across settings. Learning contexts are particularly vulnerable to these issues both in terms of what forms of AI are used in classrooms as well as what forms are used throughout the education system. The algorithms driving AI are fundamentally shaping not only our experiences in the world, but also our access to experiences in the world.

As a result, designers of AI are inherently able to reify their own cultural, historical, and political knowledges and values throughout nearly every aspect of our lives. This is made possible, in part, by the prevalent belief that technology is neutral and acultural. Hence, a critical question we should consider asking ourselves in this conversation is “AI for whom?” If the future of AI is meant to be “for all,” then this value should be reflected in policies and practice around not only the implementation of AI in learning contexts, but also the design of AI itself. It is therefore necessary to disrupt the dominant notions of who designs technologies, especially AI, away from the tech industry and education systems to the local communities who most engage with and are impacted by these technologies.

Community-based and community-driven approaches to future forms of AI could be one strategy to expanding the values and knowledges that AI embodies and supports. As an illustrative example, we share insights and findings from a multi-year critical ethnography (Madison, 2011) in a partnership project employing a community-driven design research process (Authors, 2021) between a university and a Tribal Nation. This approach to design inherently centers the goal of sustaining, revitalizing, and preserving cultural knowledge. Specifically, we outline our community-driven design process of creating learning guides as representations of cultural knowledge.

As a team of Indigenous and non-Indigenous researchers, educators, and designers, we collaboratively and iteratively analyzed design artifacts, meeting and fieldnotes, and team interviews to (re)interpret the foundational practices grounding our design process. Practices include: exchange of cultural and design knowledge, iterative interaction and discussion, transparent decision making, and a rhythm that supports reciprocity. These practices embody values that are historically disregarded by designers of learning and technology, especially in AI contexts. We argue that grounding design processes for AI in a stance that aligns with the practices and values that emerged from our community-driven design research process will expand who designs AI and, thus, for whom AI is.

Authors

• Breanne K. Litts, Utah State University
  Presenting Author

• Kenden Quayle, Utah State University
  Non-Presenting Author

• Rios Pacheco, Northwestern Band of Shoshone Nation
  Non-Presenting Author