Critical AI Approaches in the ELA Classroom

• In Event: Tinkering With/Tripping It Up: Youth Shift, Interrupt, and Challenge Conceptualizations Around Their AI Use.

Abstract

Purpose
While learning scientists have called for the examination of sociopolitical and ethical issues in computing education (e.g., Authors, 2023), such examinations remain marked primarily for students on CS pathways. Seeking to expand the purposes and reach of AI education, our researcher-practitioner team examined how learning designs in secondary English language arts (ELA) contexts can influence the types of access to rigorous, humanizing, and interdisciplinary learning about AI that students experience.

Perspectives
We draw on learning sciences perspectives that argue for reimagined disciplinary learning and the necessity of onto-epistemic heterogeneity for equitable, just forms of education (e.g., Warren et al., 2020). Onto-epistemic heterogeneity reframes disciplines through three lenses: multiplicity, horizontality, and dialogicality. The interdisciplinary endeavor of AI education in ELA engages these sensibilities by inviting stakeholders to imagine new ways to connect to this dimension of computing education in and beyond school.

Methods
We focus on two phases of a social design-based experiment (Gutiérrez, 2016) that explored culturally and linguistically diverse ELA classrooms as sites for interdisciplinary computing education (Author, 2024). In Phase 1, we observed ELA classes weekly to understand classroom life and to determine students’ baseline understandings of AI-driven sociotechnical issues. Phase 1 data sources: class observations, teacher interview, student work, class documents, and pre-unit surveys (n=55). Phase 2 focused on iterative design and implementation of a four-month unit, during which our team collaborated to develop students’ understandings of AI’s social impacts through meaningful disciplinary texts and practices. Phase 2 data sources: class observations, teacher interview, design/reflection sessions, student work, class documents, focus groups, and post-unit surveys (n=52).
Analyses during design work were concurrent with data collection. We examined student data to assess movement toward learning goals and design subsequent lessons. To construct a design case, we devised inductive codes from multiple coding cycles and deductive codes drawn from literature, theory, and SDBE design principles. Pre- and post-surveys were analyzed for shifts in students’ understandings of sociotechnical issues. Field notes, student work, and discussions were analyzed thematically for students’ responses to the lessons and for their understandings of AI’s social impacts over time.

Findings
Valued ELA tools and practices afforded multiple, extended opportunities for students to examine complex issues related to intersections of AI, race, gender, age, and mental health. Core ELA reading, writing, and discussion practices supported students in exploring and reflecting on social and ethical impacts of AI.
Findings also suggest that interdisciplinary and humanizing AI education require a strong relational foundation. Over 50% of students articulated more critical stances toward AI that they connected to their own experiences and anticipated futures. Student surveys and discussions emphasized the importance of a safe space to explore challenging and personal topics related to AI.

Significance
This study contributes to K-12 AI education research by attending to interdisciplinary learning designs that nurture young people’s critical AI literacies and in ways that reimagine ELA and computing education. The study also attends to relationality as a vital dimension of humanizing AI education and offers in-practice examples.

Authors

• Jennifer Higgs, University of California - Davis

• Maria Kaimana, envision academy