Co-Designing AI technologies with Youth: Design Choices that Reify Inequities or Envision Expansive Possibilities 

• In Event: Urban AI Unlocked: Transforming District Systems for Equitable AI Integration

Abstract

Objectives

Digital learning technologies, particularly AI, are often developed for schools without genuine engagement with students, teachers, and families. Such approaches to development limit technologies to the existing constraints of schooling, often exacerbating existing inequities (Author, 2015; Selwyn, 2013). This paper details the expansive possibilities of an extended, multi-year co-design process for developing AI with high school aged youth.

Perspective(s)

Our engagement with young people was a part of a larger AI in education institute, which included educators, educational researchers, computer scientists, and human computer interaction researchers. Our extended, interactive engagement with youth in our design process was grounded in a responsible innovation framework (Stilgoe et al., 2013) and utilized co-design (Penuel et al., 2007) and participatory design (Bang & Vossoughi, 2016) approaches.

Methods

Discussions during the interactive workshop were recorded and analyzed along with participants’ artifacts. Each discussion was transcribed, coded, and analyzed in an analytic memo. We analyzed the data for youth participants’ stances that (a) justified the institutional structuring of relations (Castoriadis, 1975) and (b) centered social dreaming (Espinoza, 2008), where they re-imagined schools as centering relationality and care. These stances were considered within the context of the workshop design choices and activities in which they were jointly produced.

Data sources

With the intention of creating space for students of color in large, urban school districts to address their hopes, dreams, and concerns about the use of Artificial Intelligence (AI) based technologies in their own schools, our multi-day interactive workshop brought together students, educators, educational researchers, and computer scientists. We sought to support high-school aged youth to envision the forms of collaboration they desired in schools, and to propose how AI may help to make those ideal collaborations a reality. The workshops included opportunities for the youth participants to experience out-of-school, real-world collaborations that formed the basis for envisioning short-term and long-term transformative possibilities for the institution of schooling.





Results

The design choices and activities of our workshops afforded and constrained possibilities for AI tools that disrupted or reproduced inequitable patterns in schools. For instance, we observed features that led students to “design for the other”—reifying problematic, taken-for-granted, static hierarchies of motivated and unmotivated students. Other approaches, however, enabled students to envision and design tools that centered collaboration, creativity, care, and compassion. Given the structure of our institute, we were able to develop a working model of their proposed design. Our findings point to design features for engaging youth in co-creating AI technologies that arise from their real, local needs and hopes and are explicitly grounded in commitments of equity.

Significance

Including potential users in the design of educational technologies is not sufficient and may even exacerbate existing inequities. Findings from our co-design activities shed light on organizing these spaces to allow participants to examine ideologies and practices that reproduce inequities and envision new possibilities as they design AI technologies for their own learning.

Authors

• Thomas M. Philip, University of California - Berkeley

• Michael Alan Chang, Boston University