Understanding K-12 Teachers' Needs for AI Education: A Survey-Based Study

Nazan Bautista, John Femiani, Daniela InclezanAAAI 2025 · DOI

Abstract

This study investigates the gap between expert-led AI training and the practical needs of K-12 educators. Through a mixed-methods survey of 34 teachers, we find that while enthusiasm for AI is high, confidence remains low. Our findings suggest that professional development (PD) should prioritize hands-on, collaborative workshops over theoretical or passive formats, as interest—rather than simple familiarity—is the primary driver of teacher confidence.

Introduction

Since the emergence of generative AI tools like ChatGPT, artificial intelligence has transitioned from a specialized computer science topic to a pervasive classroom reality. However, research has largely focused on student outcomes, leaving teachers—the primary gatekeepers of classroom technology—without adequate support. Current professional development (PD) programs are often "expert-driven," focusing on theoretical knowledge rather than the practical classroom applications teachers require. This study employs Teacher Agency Theory to explore how educators can be empowered as active participants in their own professional growth.

Methodology

The researchers utilized a mixed-methods approach to capture both quantitative trends and qualitative nuances:

  • Survey Instrument: A six-part survey covering demographics, AI knowledge, interest levels, challenges, and PD preferences.
  • Participants: 34 K-12 teachers across elementary (17.6%), middle (32.4%), and high school (53%) levels. The sample included a significant portion of veteran teachers with 11–20 years of experience.
  • Analysis: Descriptive statistics were used for general trends, while inferential statistics (ANOVA, Chi-Square, T-tests) identified correlations. Qualitative feedback was coded for thematic analysis.

Key Insights: The Familiarity-Confidence Gap

A critical finding of this study is the disconnect between knowing about AI and feeling capable of teaching it. While increased familiarity with AI correlates moderately with interest ($r = 0.40$), it has a weak correlation with confidence ($r = 0.22$).

Correlation between AI familiarity, interest, and confidence scores.

The data reveals that interest is the true gateway to confidence ($r = 0.81$). This suggests that PD should focus on sparking curiosity and demonstrating classroom relevance rather than merely providing technical definitions. Furthermore, teachers expressed a statistically significant preference for practical, hands-on training ($p = 0.0019$), emphasizing a desire to use AI for lesson planning and assessment rather than learning about underlying neural network architectures.

Professional Development Preferences

When asked about their preferred learning formats, teachers overwhelmingly chose active, social environments over passive ones.

Teacher preferences for different types of professional development.

Workshops (15 respondents) and peer collaborations (14) were the most favored, while webinars (6) were the least popular. This indicates that "passive" digital learning is failing to meet the needs of this demographic. Additionally, a one-way ANOVA showed significant differences in interest levels across grades ($p = 0.0024$), highlighting the need for tailored training rather than "one-size-fits-all" district mandates.

Institutional Barriers and Future Directions

Beyond technical literacy, teachers identified systemic hurdles including a lack of administrative support, insufficient time for training, and data privacy concerns.

While the small sample size ($N=34$) and potential selection bias are limitations, this study provides a roadmap for future AI education. The authors advocate for collaborations between AI experts and teacher educators to develop user-friendly tools that prioritize the educator's workflow. Future research will expand the sample size to better understand how needs vary across specific subject areas and grade levels.

Summary updated

BibTeX
@inproceedings{bautista2025understanding,
  author = {Nazan Bautista and John Femiani and Daniela Inclezan},
  title = {{Understanding K-12 Teachers' Needs for AI Education: A Survey-Based Study}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2025},
  doi = {10.1609/aaai.v39i28.35183},
}
Figure