Abstract:

Educator perspectives are critical determinants of effective technology adoption in schools. This research paper presents findings from a multi-stage investigation of K-12 teachers’ perspectives on the use of artificial intelligence in education (AIED) and its role in professional practice. Through a longitudinal, participatory approach combining literature reviews, surveys, interviews, focus groups, and co-design workshops, we examine what teachers perceive as the benefits and concerns of AI tools for classroom use, and how these perspectives evolve across different research contexts.

Our research progressed through five distinct phases. First, we conducted a comprehensive literature review employing systematic content analysis of 516 peer-reviewed articles, relevant reports, and conference proceedings published in the last decade, examining AI benefits, concerns, terminology, applications, and implementation contexts in K-12 settings. This groundwork informed the development of all subsequent research instruments, particularly surveys and interview guides. Phase Two involved pilot interviews with 4 teachers using pre-interview surveys and semi-structured protocols to explore initial perspectives on AI in education. These findings informed Phase Three: 3 exploration focus groups (N=9) with revised pre-session surveys and facilitated discussion to identify broader themes and concerns across diverse school contexts and teacher experience levels.

Phase Four consisted of in-person co-design focus groups (N=13) in which teachers provided detailed feedback on the proposed implementation of a GenAI support tool in their classrooms. Using mockup validation, desirability testing, and scenario brainstorming, participants moved beyond abstract discussion to concrete design engagement. Teachers initially prioritized personalized feedback as a primary benefit of AI tool use, yet when presented with specific use cases involving student data, data privacy/security emerged as a foremost concern, particularly in classroom-based scenarios regarding analyzing students’ emotional self-assessments and preparing for parent-teacher conferences.

Phase Five involved virtual co-design focus groups (N=16) in which participants generated possible use cases for the envisioned AI tool, based on a series of detailed, narrative, “What if” classroom-based scenarios, then narrowed down generated ideas to those deemed most valuable, identifying implementation priorities. This final phase revealed a notable shift: classroom-wide data analytics and personalized feedback dominated the priorities over content generation and administrative task automation, demonstrating that contextualized design scenarios yield different insights than abstract inquiry.

Our findings contribute to AIED by demonstrating that iterative, multi-method engagement with teachers reveals deeper, more nuanced perspectives and actionable insights than single-point data collection. We conclude with recommendations for responsible AIED adoption in K-12 settings, emphasizing that AIED tools, to be embraced and well utilized by educators, must center their needs, concerns, and voices throughout the design process, not as a final validation step or one-time consultation. This research illustrates both the necessity and complexity of teacher-centered AIED tool design and development.

Keywords:

Artificial intelligence in education (AIED), Participatory design, Technology adoption, Longitudinal mixed-methods research.