K-12 teachers strategically select AI tools based on their specific subject areas and instructional goals, demonstrating both enthusiasm for AI's potential and thoughtful concerns about accuracy, equity, and over-reliance.
Objective: This study examined how 227 K-12 content-area teachers (teaching mathematics, English Language Arts, social studies, and economics) used AI tools during graduate coursework, focusing on which tools they preferred, what instructional goals these tools supported, and what challenges emerged during integration. The researchers sought to move beyond general examinations of AI adoption by providing nuanced, subject-specific insights into how teachers across different disciplines engage with educational AI technologies.
Methods: The researchers employed a qualitative-dominant mixed-methods design within a graduate-level Curriculum and Instruction program at a large public university in the southeastern United States. Participants were recruited through convenience sampling from enrolled graduate students who were also practicing K-12 teachers. As part of a course assignment, teachers explored various AI features on Teacherserver.com, a free platform created by the researchers offering over 1,000 AI-powered applications for educators. Teachers used these tools to design lessons, generate assessments, and tailor instruction for their specific content areas.
Data collection occurred through open-ended online surveys distributed via Qualtrics, where participants identified their five favorite tools, explained their choices, described anticipated instructional applications, reflected on challenges, and discussed future professional learning plans. The survey was reviewed by faculty experts and piloted with four K-12 teachers before finalization. The research team analyzed qualitative data using Braun and Clarke's six-phase thematic analysis process: familiarization, coding, theme development, theme review, theme definition, and final reporting. Two researchers independently coded responses and met to reconcile discrepancies, achieving over 85% initial agreement. The analysis was framed through the Technological Pedagogical Content Knowledge (TPACK) framework, which emphasizes the intersections of teachers' technological, pedagogical, and content knowledge.
Key Findings:
The study revealed clear subject-specific patterns in AI tool selection. Mathematics and algebra teachers (41.4% of participants) favored tools supporting differentiation and problem-solving, such as the Math Word Problem Generator and Problem-Solving Strategy Generator. One teacher noted the tools helped manage differentiation needs across students ranging from second to mid-seventh grade levels. Economics teachers (7.0%) preferred real-world application tools like the Budget Planning Tool and Market Competition Simulator, valuing how these made abstract concepts concrete. ELA teachers (24.7%) selected tools supporting writing and vocabulary development, including the Vocabulary Enhancer and Writing Assessment Generator, particularly appreciating support for compound sentence construction. Social studies teachers (26.9%) emphasized tools promoting critical thinking and reading comprehension, such as the Text-to-Question Converter and Inference Builder.
Teachers identified three major challenge categories: content alignment and accuracy (concerns about reliability and standards alignment), teacher preparedness and confidence (limited prior training leading to uncertainty), and equitable access and over-reliance (inconsistent device/internet access and worries about student dependence). Despite these concerns, teachers demonstrated strong interest in learning more about AI through professional development, peer collaboration, and self-directed exploration via blogs, social media, and online communities.
Implications: The findings emphasize that effective AI integration in education requires subject-specific tool design and professional development rather than one-size-fits-all approaches. The TPACK framework illuminates how teachers strategically negotiate relationships among technology, pedagogy, and content knowledge when selecting AI tools. This study contributes to broader conversations about preparing educators for contextually meaningful AI use by highlighting the importance of aligning technological innovations with disciplinary instructional practices. The research suggests that successful AI adoption depends not merely on tool availability but on how teachers develop intersecting competencies across technological, pedagogical, and content domains.
The study also underscores the need for ongoing, structured professional development that goes beyond technical training to include critical evaluation of AI outputs, ethical considerations around bias and equity, and curriculum alignment strategies. Teachers' willingness to learn, combined with their cautious approach, indicates readiness for well-designed support systems that preserve educator agency while leveraging AI capabilities.
Limitations: The study has several important limitations affecting generalizability. All participants were graduate students enrolled in a single university program, representing convenience sampling that may not reflect the broader K-12 teacher population. These graduate students likely had higher motivation, professional engagement, and access to professional development than typical practicing teachers. Data collection relied entirely on self-reported survey responses, introducing potential social desirability bias and incomplete recall. Teachers may have emphasized positive experiences while downplaying difficulties. Additionally, the relatively small number of economics teachers (n=16) limits the strength of conclusions for that discipline. The study did not include classroom observations or longitudinal data tracking how AI use evolved over time.
Future Directions: The researchers recommend several directions for future investigation. Longitudinal studies should explore how teacher perceptions and practices evolve as classroom experience with AI tools increases. Content-specific research should examine unique disciplinary practices—such as AI-assisted mathematical problem-solving versus AI-enhanced writing feedback—to provide tailored guidance for tool development and professional learning. Comparative analyses across elementary, middle, and high school levels could reveal distinct opportunities or challenges at different educational stages. Multi-site or cross-institutional studies including both graduate students and practicing teachers outside coursework contexts would help assess whether findings transfer across diverse professional settings. Finally, research should examine how policy frameworks, institutional supports, and infrastructure investments influence equitable AI integration, particularly in under-resourced schools where the digital divide may be most pronounced.
Title and Authors: "Exploring K-12 content-area teachers' preferences and challenges in using AI tools in graduate coursework" by Aslihan Unal (Georgia Southern University) and Zafer Unal (University of South Florida).
Published On: The article was received on May 21, 2025, accepted on August 20, 2025, and published online on September 30, 2025.
Published By: AI and Ethics, published by Springer Nature Switzerland AG. The article DOI is https://doi.org/10.1007/s43681-025-00833-2.