2025 ASEE Annual Conference & Exposition

This NSF RAPID project (award number 2333393) addresses the recent influx of generative artificial intelligence (AI) tools (e.g., ChatGPT, Khanmigo, Midjourney, etc.) available to the public. Many of these tools have potential uses and misuses for educational purposes and can have applications as new and existing tools in classrooms. It is important for both researchers and AI tool developers understand kindergarten through twelfth grade (K-12) teachers’ perceptions of AI use in the classroom. Namely, how do K-12 teachers perceive AI tools in education and the impact of AI on the workforce?

Our team developed a novel survey developed from Biesta and Tedder’s (2007) ecological agency model, Bronfenbrenner’s (2000) ecological systems theory, and preexisting teacher agency subscales (Liu, et. al; 2016). This survey included 44 Likert-type questions, three open-response questions, and 14 demographic questions. We collected 1,000 complete, unique, and verified responses from K-12 teachers in the United States and territories. Analysis included descriptive statistics, open-response thematic analysis, exploratory factor analysis, linear regression modeling, and comparative analysis across subpopulations.

In general, teachers endorsed items suggesting AI tools could be valuable (e.g., 78% agreed that AI tools could help support them with key challenges), however they also endorsed items suggesting ethical and learning concerns with AI tool use (e.g., 85% had ethical concerns about students’ AI use). Open response results indicated that teachers saw potential AI applications for creating lesson materials, supporting students with different learning needs, grading, managing large data, detecting cheating, and communicating via writing (e.g., routine communicating with parents). Teacher concerns included student cheating, teacher use, not being developmentally appropriate for younger grades, unreliability, and ethical concerns. Opinions about AI tools being helpful in the classroom were statistically different between high, middle, and elementary school, with high school teachers expressing more belief of AI tool utility in the classroom. Linear regression modeling described explicit school support (e.g., allowing or encouraging AI tool use, technology resources in schools) as predicting an increase in the impact of teacher AI tool use across all grade levels, with the impact being the strongest for elementary school teachers.

Our results suggest that teachers’ agency for AI tool use, including their feelings of about their own time, ability, and values, is salient in their willingness to use AI in their classrooms. Additionally, social support among teachers and from school leadership can increase teachers’ willingness to use AI tools. These results support providing training that specifies grade-appropriate use to all K-12 teachers about AI tools in the classroom.

Citations

Biesta, G. & Tedder, M., “Agency and learning in the life course: Towards an ecological perspective,” Stud. Educ. Adults, vol. 39, no. 2, pp. 132–149, Sep. 2007, doi: 10.1080/02660830.2007.11661545.

Bronfenbrenner, U. (2000). Ecological systems theory. American Psychological Association.

Liu, S., Hallinger, P., & Feng, D. (2016). Supporting the professional learning of teachers in China: Does principal leadership make a difference?. Teaching and teacher education, 59, 79-91.