Abstract
Teaching AI ethics in K-12 settings is increasingly vital but faces a significant expertise gap between technical and humanities disciplines. This study examines how four secondary teachers (ELA and CS) adapted a modular AI ethics curriculum using the TPACK framework. Findings suggest that teachers successfully customize education by leveraging their existing domain expertise and local context, highlighting the value of flexible, co-designed curricula over rigid, scripted models.
Introduction
As Artificial Intelligence (AI) becomes ubiquitous, the necessity of integrating AI ethics into K-12 education has reached a critical point. However, this interdisciplinary subject often falls into an "expertise gap": it requires technical knowledge that humanities teachers may lack, and a deep engagement with complex social issues—such as race, gender, and class—that STEM teachers may find challenging to facilitate.
This paper explores how secondary school teachers from diverse disciplines (English Language Arts and Computer Science) adapt a modular AI ethics curriculum to fit their specific classroom contexts. By moving away from "curriculum fidelity," the researchers demonstrate that teachers' local knowledge and professional expertise are essential for effective AI education.
Theoretical Framework: TPACK
The study utilizes the Technological Pedagogical Content Knowledge (TPACK) framework to categorize how teachers integrate technology, pedagogy, and content knowledge. TPACK posits that effective teaching with technology requires an understanding of how these three domains overlap within specific educational contexts.
Methods
The researchers employed a collaborative co-design approach with four secondary teachers: two 9th-grade ELA teachers, one 8th-grade CS teacher, and one 9th-grade Cybersecurity teacher. The curriculum used a modular, project-based learning (PBL) framework involving original short stories, hands-on AI tools (e.g., Google’s Teachable Machine), and digital media design tasks.
Qualitative data was collected through modified lesson plans, classroom observations, and semi-structured interviews to understand the rationales behind teachers' adaptations.
Case Studies and Implementation
The study analyzed four distinct implementations across urban, suburban, and semi-rural schools. The following table summarizes the participants and the specific ways they adapted the curriculum to their students' needs.
Expertise-Driven Customization
Teachers modified the curriculum most heavily in areas where they were already experts. For example, an ELA teacher with a Ph.D. in Literacy replaced standard modules with a graphic novel version of Frankenstein to explore "monstrosity" in AI. In contrast, a CS teacher integrated physical computing (micro:bits) into the final ethics project.
The "Support" Threshold
Teachers relied on the research team for components furthest from their expertise. ELA teachers used technical AI modules "as-is," while CS teachers sought guidance on leading sensitive discussions about social bias.
Key Insights
- Contextual Sensitivity: Adaptations were often driven by local culture. A teacher in a conservative semi-rural area focused on community-building, while a teacher in a social-justice-oriented urban school focused on structural inequality and racial bias.
- PBL as a Bridge: Project-based learning allowed both STEM and Literacy teachers to meet their respective state standards while tackling core AI ethics concepts.
- The TPACK Shift: Researchers observed that teachers transformed their existing pedagogical strategies to accommodate the interdisciplinary nature of AI ethics.
Conclusion
The study found that teachers can provide high-quality AI ethics instruction by "remixing" curricula to align with their strengths. While the reliance on researcher support highlights potential scalability issues, the success of these adaptations suggests that interdisciplinary co-teaching could be a powerful model for future AI education.
Summary updated