Utilizing Co-Design to Introduce Technological and Algorithmic Bias in Middle School Computer Science Lessons (Poster 8)

• In Event: Perspectives on Promoting Algorithmic Justice in Artificial Intelligence and Machine Learning Education in K–12

Abstract

Objectives or purposes: Despite assumptions that technology and algorithms are unbiased (Pew Research Center, 2018), research has demonstrated the bias and prejudice inherent in, and caused by, technologies (e.g., Benjamin, 2019; Obermeyer et al., 2019). As such, in computer science classes students need to learn not only how to program, but also the social justice implications of the technologies they use and create. This design study brought together teachers and students to develop a curriculum to teach about both algorithmic and technological biases and computer science skills.
Theoretical framework: This work is framed through participatory design. Participatory design centers user voices (Muller & Kuhn, 1993) with the goal of equalizing power structures (Bjerknes & Bratteteig, 1995; Bødker et al., 2000) to democratize design (Björgvinsson et al., 2010; Iversen et al., 2004) and promote user voices and identities (Coenraad et al., 2019). Within education it has been utilized to center teacher and student voices and develop novel learning experiences and curricula (Coenraad et al., 2021; Penuel et al., 2007; Pinkard et al., 2017; Proctor & Blikstein, 2019).
Methods and Data Sources: Using participatory design methods, we collaborated with eight youth (ages 8 - 13; 5 boys, 3 girls; 8 Black/African-American) over six sessions and two teachers (2 men; 1 Black/African-American, 1 Latinx/Latine) over eight meetings to understand how youth conceptualize technological and algorithmic bias and how this understanding informs co-design of a curriculum. We present an analysis of the first unit from the Talking Techquity curriculum (Techquity is a portmanteau of technology and equity) to highlight the integration of youth and teacher design ideas. Our analysis identified the origins of components of the unit plans (e.g., theme, activities) using field notes, transcripts, and a design log and open coded the component origins to identify trends in methods for integration.
Results: The sequence, topics, and pedagogical strategies in the first unit of the Talking Techquity curriculum were drawn from ideation and iteration at the co-design sessions. The unit focuses on the computing concept animation using loops and the Techquity topic algorithmic bias in image search results. It follows a four-lesson sequence building on the Use, Modify, Create pedagogical strategy (Lee et al., 2011) with an initial lesson introducing algorithmic bias in image search results. The unit topic of algorithmic bias was drawn from youth and teacher examples of how technology can negatively impact youth’s lives. Additionally, the unit utilizes pedagogical strategies, such as emphasizing student participation through discussion, ensuring students are provided with clear definitions, and providing experiential learning opportunities, based on designer ideas.
Significance: This study highlights how both youth and teacher voice can be integrated into a curriculum and how technological and algorithmic bias can be taught alongside computer science skills. Youth are aware of the positive and negative impacts of technology (Coenraad, 2022). Building on this understanding, their preferred learning methods, and teacher expertise provides opportunities for students to become even more attuned to how technology can impact them and prepares them for our increasingly computational world.

Authors

• Merijke Coenraad, Digital Promise Global

• David Weintrop, University of Maryland