An Artificial Tool or Natural Support? Teachers’ Utilization of AI in Early Literacy Instruction

• In Event: Computational Playgrounds: Foundations and Futures of AI, Play, and Pedagogy in Early Childhood

Abstract

Objectives or purposes
The integration of Artificial Intelligence (AI) in early childhood education is rapidly expanding, particularly through the use of generative AI tools (e.g., ChatGPT) and interactive technologies (e.g., educational robots). While current discourse emphasizes the technical affordances of AI, it often overlooks the critical perspectives and pedagogical agency of early childhood educators who mediate how AI is meaningfully implemented in classrooms. This mixed-methods study investigates early childhood teachers’ perceptions and classroom applications of AI, specifically exploring whether AI is perceived as an “artificial tool” or a “natural support” in early literacy instruction. The study aims to (1) examine in-service teachers’ perceptions of AI integration in classroom teaching, and (2) observe how AI is incorporated into typical classroom practices to inform future models that position AI as a collaborative partner in early literacy instruction.

Perspective(s) or theoretical framework.
Framed by the metaliteracy framework (Jacobson & Mackey, 2017), the study conceptualizes AI integration across four interrelated domains of literacy learning: behavioral (technical skills), cognitive (knowledge acquisition), affective (emotional responses), and metacognitive (reflective awareness). This framework aligns with the National Association for the Education of Young Children (NAEYC) recommendations for holistic child development and echoes the K–12 Computer Science Framework’s emphasis on developmentally appropriate computational thinking in early childhood.
Methods, techniques, or modes of inquiry
Twenty-two preschool teachers (all female; 13 African American, 8 Caucasian, and 1 Asian American) completed an AI literacy survey and individual interviews. Among these teachers, six teachers participated in an extended phase of the study; they were observed monthly over three sessions and interviewed afterward. These teachers were provided with tablets equipped with Gemini and ChatGPT and were encouraged to use AI in their existing classroom activities at their own discretion. Observations were conducted using the Early Language and Literacy Classroom Observation Pre-K (Smith et al., 2002) and a self-developed AI literacy observation tool. Each classroom observation lasted for 2-3 hours.
Results
Findings revealed that while all teachers had basic technological competence for independently using AI, only four used AI for professional needs (e.g., lesson planning), and only three teachers—those with prior personal AI experience—applying AI for instructional or classroom management tasks (e.g., searching information). Although attitudes toward AI were generally positive, structural and cultural barriers, including restrictive program policies on screen time and concerns about parental acceptance, constrained teachers’ AI utilization. Most teachers learned about AI informally, and several reported that their formal training either neglected or discouraged AI use.
Classroom observations indicated that low-level AI use (e.g., timers, music) had minimal instructional impact, while mid- and high-level utilization—such as using AI to search for information during book reading or co-create narratives based on children's input (see Figures 3-1 and 3-2) significantly enhanced teacher-child extended conversations and children’s engagement. The higher-level utilizations were more strongly associated with children’s age and teachers’ instructional efficacy than with teachers’ technical computing skills. These findings suggest that AI, when naturally embedded into instruction, holds promise as a pedagogical partner in meaningful early literacy instruction.

Authors

• Chenyi Zhang, Georgia State University

• Lauren Margulieux, Georgia State University

• Nooshin Haddadian, Georgia State University