A short, project-based AI course significantly increases high school students’ AI knowledge, while AI anxiety remains stable but can positively support learning gains.

Objective:
The purpose of this study was to examine how a six-lesson artificial intelligence (AI) course, grounded in the AI4K12 Five Big Ideas framework and delivered through project-based learning (PBL), influenced high school students’ AI knowledge, AI anxiety, and engagement. The research also aimed to determine whether students’ pre-existing anxieties toward AI were associated with their learning outcomes or levels of participation during collaborative project activities.

Methods:
The study used a mixed-methods design involving both quantitative and qualitative data. Participants were high school Teacher Cadet students enrolled at a vocational school. The instructional intervention consisted of synchronous lessons, blended discussions, collaborative group projects, and final project presentations. Quantitative measures included pre- and post-tests of AI knowledge, the Artificial Intelligence Anxiety Scale (AIAS), the Comprehensive Assessment of Team Member Effectiveness (CATME) for engagement, and rubric-based project evaluations. Qualitative data consisted of classroom observations, video-recorded reflections, and written responses. Statistical analyses and thematic coding were used to evaluate changes in knowledge, affect, and engagement.

Key Findings:

• Students demonstrated significant growth in AI knowledge, supported by higher post-test scores and richer conceptual explanations in their qualitative reflections.

• AI anxiety did not significantly change overall following the course, and subscales—including Configuration Anxiety, Job Replacement Anxiety, and Sociotechnical Blindness—remained stable.

• Learning Anxiety was positively correlated with knowledge gains, meaning students with higher initial worry about learning AI tended to improve more.

• AI anxiety did not predict engagement levels during PBL activities.

• Engagement scores showed a moderate positive correlation with final project performance, indicating that more active collaborators produced higher-quality work.

• Students could clearly distinguish between automation and AI but struggled with explaining complex or layered AI systems.

Implications:
This study demonstrates that high school students can effectively learn foundational AI concepts within a short instructional period when PBL and the AI4K12 framework are combined. It highlights the value of hands-on, collaborative learning environments and the importance of scaffolding to help students understand complex AI mechanisms. The findings also show that anxiety toward AI is not inherently harmful; in some cases, learning-related anxiety may motivate students to engage more deeply. The study contributes to the growing field of AI literacy research by clarifying how cognitive and affective factors interact during AI learning experiences in K–12 settings.

Limitations:
The study involved a small sample size, limiting generalizability. The short duration of the intervention may not have been sufficient to shift deeply held anxieties about AI. Gender imbalances within the group may also have influenced engagement dynamics. Additionally, engagement measures relied on peer reports and observational judgments, which may introduce subjectivity.

Future Directions:
The study recommends expanding research to larger and more diverse samples, implementing longer-term AI instructional programs, and examining incremental learning across each of the AI4K12 Big Ideas. Future research should include more detailed measures of affective changes, closer analysis of group collaboration dynamics, and stronger integration of ethical, social, and real-world AI applications. Continued work is needed to understand how emotional factors like anxiety influence both learning and engagement over longer periods of time.

Title and Authors: A Mixed-Methods Exploration of Knowledge and Perceptions of Artificial Intelligence Among High School Students by Erin M. Bosarge.

Published On: December 2025.

Published By: University of South Alabama (Doctoral Dissertation).