Teachers’ Competence, Self-Efficacy, and Their Attitudes Toward Generative AI in Education: A Correlational Study

The rapid integration of Generative Artificial Intelligence (GenAI) into educational settings presents a complex landscape of opportunities and challenges. A critical factor for the successful implementation of these powerful tools lies in teacher adoption. However, a significant gap has persisted in understanding the psychological drivers that shape teachers’ attitudes toward GenAI. A recent study, published in Frontiers in Psychology, sheds crucial light on this matter, revealing that while digital competence is important, a teacher’s self-efficacy—their confidence in their ability to effectively use GenAI—emerges as the most potent predictor of their positive attitude towards its integration in the classroom.

The research, conducted with 352 K-12 teachers in a large urban school district in Indonesia, employed a quantitative, cross-sectional survey design. It meticulously examined the relationships between teachers’ general digital competence (GDC), their AI-specific competence (AISC), their self-efficacy for AI integration (TSE-AI), and their attitudes toward Generative AI in Education (ATGAI-E). The findings underscore a nuanced understanding of teacher preparedness for the AI era, suggesting that fostering confidence may be as, if not more, critical than solely enhancing technical skills.

The Evolving Landscape of AI in Education

The advent of GenAI tools like ChatGPT has propelled artificial intelligence from a theoretical concept to a tangible reality within classrooms worldwide. These tools, capable of generating human-like text, code, and even images, hold immense potential to revolutionize pedagogy. Applications range from personalized student tutoring and automated content creation to providing instant feedback and streamlining administrative tasks. However, the history of educational technology is replete with innovations that failed to achieve widespread adoption due to a lack of teacher buy-in. As educational institutions grapple with developing policies and strategies for GenAI integration, understanding the psychological barriers and enablers for teachers is paramount.

This study addresses a critical need for empirical evidence by focusing on the interplay of competence and self-efficacy in shaping teacher attitudes. Drawing upon established theories like the Technology Acceptance Model (TAM) and the TPACK framework, the research posits that teachers’ perceptions of usefulness and ease of use—key determinants of technology adoption—are significantly influenced by their skills and confidence levels. Specifically, the study differentiates between general digital competence and AI-specific skills, recognizing that the latter requires a deeper understanding of AI functionalities, prompt engineering, and ethical considerations.

Unpacking the Factors Influencing Teacher Attitudes

The study’s methodology involved administering an online questionnaire that assessed four key constructs using a 5-point Likert scale. The psychometric properties of the scales were rigorously validated, demonstrating high internal consistency and acceptable model fit. Pearson correlations revealed strong, positive relationships among all variables. Attitudes toward GenAI (ATGAI-E) showed the strongest correlation with Teacher Self-Efficacy for AI Integration (TSE-AI), with a coefficient of 0.75 (p < 0.001). This was followed by AI-Specific Competence (AISC) at 0.69 (p < 0.001) and General Digital Competence (GDC) at 0.58 (p < 0.001). These findings initially suggest that as teachers’ skills and confidence increase, so does their positive outlook on GenAI.

However, the true power of these factors was illuminated through hierarchical multiple regression analysis. This advanced statistical technique allowed researchers to determine the unique predictive contribution of each variable while controlling for demographic factors such as teaching experience, grade level, and subject area. The analysis revealed that the initial model, including only demographic variables, explained a mere 2.6% of the variance in teachers’ attitudes.

The introduction of the psychological predictors in the second step dramatically shifted the explanatory power of the model. The combined psychological factors accounted for an additional 61.3% of the variance, leading to a robust overall model that explained 62% of the variance in teachers’ attitudes toward GenAI. Crucially, Teacher Self-Efficacy for AI Integration (TSE-AI) emerged as the most significant predictor, with a standardized beta coefficient of 0.48 (p < 0.001). This indicates that even after accounting for demographic characteristics and both general and AI-specific digital skills, a teacher’s confidence in their ability to integrate GenAI remained the most powerful driver of their positive attitude.

AI-Specific Competence (AISC) also played a significant role, with a beta coefficient of 0.25 (p < 0.001), highlighting the importance of understanding the nuances of AI. General Digital Competence (GDC) contributed positively, albeit to a lesser extent, with a beta of 0.11 (p = 0.015). This hierarchical breakdown clearly demonstrates that while foundational digital skills and specific AI knowledge are valuable, the psychological empowerment derived from self-efficacy is the lynchpin for fostering positive teacher attitudes towards GenAI.

Deconstructing Teacher Self-Efficacy

The study further delves into the multifaceted nature of teacher self-efficacy for AI integration. While the overall score proved to be the strongest predictor, a more granular analysis reveals potential sub-dimensions of this construct. These might include confidence in instructional design, pedagogical application, classroom management and assessment related to AI use, and ethical and technical proficiency. This nuanced view is critical because a teacher might possess strong technical skills in one area but lack confidence in another, such as assessing AI-generated student work or guiding discussions on AI ethics. This complexity helps explain why teachers with seemingly adequate technical knowledge might still exhibit hesitancy towards adopting GenAI. Their apprehension might stem not from a general lack of ability but from specific deficits in confidence within critical aspects of their practice.

This layered understanding of self-efficacy suggests that professional development initiatives must move beyond generic training. Instead, they should aim to identify and strengthen specific areas of low confidence. For instance, a teacher may feel adept at using AI to create lesson plans but anxious about its implications for academic integrity. Targeted interventions addressing these specific concerns could be far more effective in building overall confidence and fostering a more positive attitude.

Implications for Professional Development and Educational Policy

The findings carry significant weight for educational leaders and policymakers tasked with facilitating the integration of GenAI. The paramount role of self-efficacy suggests that professional development programs must prioritize confidence-building alongside skill enhancement. This could involve:

• Experiential Learning: Providing ample opportunities for teachers to engage in hands-on practice with GenAI tools in low-stakes environments, focusing on real-world classroom challenges. Success in these practical applications can significantly bolster self-efficacy.
• Peer Modeling and Mentorship: Showcasing successful examples of GenAI integration by experienced educators can inspire confidence in others, fostering a belief that similar achievements are attainable.
• Supportive Feedback Mechanisms: Implementing systems where instructional leaders and peers offer specific, constructive feedback and encouragement can foster resilience and motivation.
• Addressing Specific Competency Gaps: While self-efficacy is key, the study also highlights the importance of AI-specific competence. Professional development should include targeted training on prompt engineering, understanding AI outputs, data privacy, algorithmic bias, and the ethical implications of AI use.

The research implies that a teacher’s journey toward embracing GenAI is deeply intertwined with their psychological readiness. A culture that minimizes stigma around uncertainty and encourages experimentation is crucial. Educational institutions that foster such an environment are more likely to see their teachers not just adopt but creatively integrate GenAI to enhance teaching and learning outcomes.

Limitations and Future Directions

While this study offers valuable insights, its cross-sectional design means it cannot definitively establish causality. Future longitudinal research tracking teachers’ development over time would provide a more robust understanding of how competence, self-efficacy, and attitudes evolve. Furthermore, the study’s reliance on a convenience sample from a single urban district limits its generalizability. Replicating these findings with more diverse populations across different geographical and socio-economic contexts is essential. The use of self-report measures also introduces the potential for social desirability bias, suggesting that future studies could benefit from incorporating objective assessments of competence and observed classroom practices. As the field of GenAI continues its rapid evolution, ongoing research will be vital to keep pace with technological advancements and their impact on the educational landscape. Qualitative studies could also offer deeper, contextualized understanding of teachers’ lived experiences with GenAI integration.

Conclusion

The integration of Generative AI into education is not merely a technological upgrade but a pedagogical paradigm shift. The success of this transition hinges significantly on the preparedness and receptivity of educators. This research unequivocally demonstrates that while foundational digital skills and specific AI knowledge are important, a teacher’s confidence in their ability to leverage these tools—their self-efficacy—is the most critical determinant of their positive attitude towards GenAI. Teachers who feel empowered and confident are more likely to be open, optimistic, and proactive in exploring how GenAI can enrich teaching and learning.

Educational leaders must recognize that fostering positive adoption goes beyond providing access to technology or basic training. It necessitates a holistic approach that prioritizes building teacher confidence through supportive, experiential professional development. By focusing on empowering teachers, encouraging experimentation within safe environments, and addressing specific concerns, educational institutions can cultivate a workforce that is not only adept at using AI tools but also resilient, creative, and forward-thinking, ready to shape the future of education in the AI era. This study contributes significantly to the understanding of teacher self-efficacy in the context of AI adoption, providing a crucial framework for guiding effective implementation strategies.