A six-week professional development course significantly increased K-12 teachers' comfort with AI, their perceptions of AI, and access to AI resources, demonstrating that structured training can effectively prepare educators to integrate artificial intelligence into their classrooms.

Objective: The main goal of this study was to understand if a professional development course affected K-12 teachers' comfort with AI, perceptions of AI, and access to resources, while also examining relationships between these variables and demographic factors that might influence AI comfort levels.

Methods: The researcher employed a single-group pretest-posttest design using secondary data from the Auburn University Biggio Center for the Enhancement of Teaching and Learning. The study involved 30 participants who completed a six-week professional development Canvas course consisting of six content modules covering AI fundamentals, ethical considerations, information literacy, AI tools for learning, teaching applications, and assessment methods. Data was collected through Qualtrics surveys administered before and after course completion, measuring three main constructs: comfort with AI, perceptions of AI, and access to AI resources. The analysis included paired-sample t-tests to compare pre- and post-test means, correlations to examine relationships between variables, independent samples t-tests for binary demographics, factorial ANOVA for multi-group demographics, and reliability testing using Cronbach's alpha for each construct.

Key Findings:

• All three constructs showed statistically significant improvements from pre-test to post-test: comfort with AI (p < .001, Cohen's d = 1.07), perceptions of AI (p = .011, d = .50), and access to resources (p = .003, d = .60).
• Participants' familiarity with AI increased significantly from 5.20 to 7.27 on a 10-point scale (p < .001, d = .74), while openness to learning remained high but unchanged due to ceiling effects.
• Strong positive correlations existed between perceptions of AI and comfort with AI both before (ρ = .538, p = .002) and after (ρ = .446, p = .014) the course.
• Significant relationships were found between access to resources and comfort with AI in both pre-test (ρ = .549, p = .002) and post-test (ρ = .525, p = .003) conditions.
• No demographic factors (gender, age, years of teaching, education level, employment status, ethnic background, school type, or Title 1 status) showed significant differences in comfort levels with AI.
• The course demonstrated excellent internal consistency reliability for comfort with AI (α = .818) and access to resources (α = .876), with good reliability for perceptions of AI (α = .761).

Implications: This research provides crucial evidence that structured professional development can effectively increase educator confidence and competence with AI technologies. The findings suggest that well-designed training programs can successfully bridge the gap between AI hesitancy and classroom implementation, regardless of teacher demographics. The strong correlations between perceptions, resource access, and comfort levels indicate that comprehensive training addressing multiple dimensions of AI literacy is essential. The study demonstrates that professional development programs should focus on building both theoretical understanding and practical access to resources to maximize teacher comfort and adoption rates. These findings support the development of standardized AI literacy curricula for K-12 educators and highlight the importance of providing accessible, high-quality training opportunities to prepare teachers for technology integration.

Limitations: The study faced several significant limitations that affect the generalizability of findings. The small sample size (N=30) limits statistical power and the ability to detect demographic differences that might exist in larger populations. The six-to-eight-hour course duration created recruitment challenges due to teachers' busy schedules, potentially limiting participation to more motivated educators. Most participants were likely innovators or early adopters of AI technology, creating selection bias that may not represent the broader teacher population, particularly those in the majority or late adopter categories. The study's reliance on self-generated identification codes resulted in some data loss when participants used different codes between surveys. Additionally, the majority of participants (93.3%) came from private schools, which may have different technology access and resources compared to public schools, limiting representativeness across diverse educational contexts.

Future Directions: The researcher recommends several important areas for future investigation, including conducting mixed-methods studies that incorporate follow-up interviews with course participants to gain deeper insights into specific comfort areas and ongoing needs. Future research should focus on recruiting participants across the full spectrum of technology adoption, particularly targeting majority and late adopters rather than just early adopters, to better understand the course's effectiveness across different comfort levels. Studies with larger sample sizes are essential to examine demographic differences more thoroughly and improve statistical power. Researchers should investigate the long-term impact of AI training by conducting longitudinal studies that track participants' actual classroom implementation over time. Additional research should explore the development and validation of standardized scales to measure technology adoption readiness and compare this framework's effectiveness for other emerging technologies like virtual and augmented reality. Finally, studies should examine how different delivery methods and course durations affect learning outcomes and participant engagement.

Title and Authors: "K-12 Educators' Comfortability and Perception of Artificial Intelligence" by Katelyn R. Nelson (Auburn University), with committee members David Marshall, Asim Ali, Ellen Hahn, Reginald Blockett, and Malti Tuttle.

Published On: August 9, 2025

Published By: Auburn University (Doctoral dissertation submitted to the Graduate Faculty in partial fulfillment of the requirements for the Degree of Doctor of Philosophy)