Tirtha Karki
This study investigated the impact of integrating AI literacy into content-area instruction for a multilingual (ML) course on in-service teachers’ Intelligent-Technological, Pedagogical, and Content Knowledge (Intelligent-TPACK), their practices of AI use in teaching, and their perspectives on the affordances and challenges of implementing AI ethically and critically in ML classrooms. It is informed by four interrelated theoretical frameworks (AI literacy, Intelligent-TPACK, sociocultural theory, and critical pedagogy), guiding AI literacy module development and research procedures. Employing a convergent mixed methods case study, I collected and analyzed data from twenty-four in-service teachers who completed an online graduate course in Spring 2025. The quantitative data include pre- and post-Qualtrics surveys consisting of 31 items on a 5-point Likert scale, while the qualitative data encompass lesson plans, lesson reflections, discussion board responses, VoiceThread discussions, semi-structured interviews, and artifacts. The quantitative data were analyzed descriptively with SPSS, whereas the qualitative data were analyzed thematically using NVivo. The findings revealed that in-service teachers’ Intelligent-TPACK scores changed positively after the course. However, they did not demonstrate knowledge and skills for intelligent tutoring, interactive learning, automated, real-time, and individualized feedback, and leadership. Results showed that teachers primarily used AI for materialistic purposes (e.g., lesson planning, rubrics, and assessment tools). However, they used AI infrequently for some prominent tasks (e.g., social interactions, language and content integration, and comprehensible input). Teachers also reported ethical and critical concerns about GenAI use (e.g., biases, lack of linguistic and cultural representations, emotionally unresponsive, and monolingual ideologies). The findings offer implications for various stakeholders.