Pathways to AI Literacy among Pre-service Teachers: Evidence from Structural Equation Modelling

• In Event: Reimagining Education in Complex Contexts: Curriculum, Community, and Teacher Capacity

Proposal

Introduction and Rationale

The rapid growth of emerging technologies, global connectivity, and the COVID-19 pandemic have accelerated the digital transformation of education, placing AI at the forefront of teaching. Effective classroom integration requires teachers to have AI literacy—the ability to understand, apply, monitor, and critically reflect on AI to enhance teaching quality and support students’ knowledge and skill development (Ai Chu et al., 2024).

As key drivers of future education, pre-service teachers’ AI literacy shapes their professional growth, classroom innovation, teaching quality, and educational equity. This has gained international attention: UNESCO’s AI Competency Framework for Teachers identifies 15 core competencies—from ethical responsibility to instructional applications—and proposes a three-stage “acquire–deepen–create” pathway (UNESCO, 2025). Similarly, China is actively promoting AI integration in teacher education through both policy and practice, emphasizing AI-supported teaching, curriculum incorporation, practicum models, and classroom assessment (Central Committee of the Communist Party of China & State Council, 2025; Ministry of Education of the People’s Republic of China, 2025).

As policies increasingly prioritize AI literacy, its cultivation among future teachers has become strategic. Yet empirical research on pre-service teachers remains limited, focusing mainly on definitions (Chiu et al., 2024; Sperling et al., 2024) or training outcomes (Zhang et al., 2024) rather than theory-driven analyses of cognition, social influence, and behavioral intention. This study takes Chinese pre-service teachers’ AI literacy as an example, integrating the Technology Acceptance Model (TAM) and Social Cognitive Theory (SCT) to examine how personal and contextual factors influence AI literacy, providing a theoretical framework for teacher education.


Model and Hypothesis Development

Based on the core constructs from Technology Acceptance Model (TAM) and Social Cognitive Theory (SCT), six key potential influential factors were identified for this study, i.e. Perceived Ease of Use, Perceived Usefulness, and Behavioral Intention to Use from TAM; Self-efficacy, Observation Learning, and Outcome Expectations from SCT. This study designed an influential factors model to examine the relationship between these six factors and pre-service teacher AI Literacy. The rationale of integrating these two models is to address the limitations of a single model in explanatory power. TAM primarily focuses on individual perceptions and acceptance intentions of technology, with little consideration of external and contextual factors. In contrast, SCT emphasizes the interplay between personal factors, behavior, and environmental influence (Chun et al., 2025). In addition, Self-efficacy and Observational Learning can serve as important exogenous explanatory variables for Perceived Ease of Use, while Outcome Expectation complements the role of Perceived Usefulness in Behavioral Intention to Use from a motivational perspective, thereby providing a more comprehensive explanation of the formation mechanism of AI literacy.

The sixteen hypotheses proposed in this study are as follows:
H1: Self-efficacy positively affects pre-service teacher AI Literacy.
H2: Self-efficacy positively affects Observation Learning.
H3: Self-efficacy positively affects Behavioral Intention to Use.
H4: Self-efficacy positively affects Perceived Ease of Use
H5: Observation Learning positively affects Perceived Ease of Use
H6: Observation Learning positively affects Perceived Usefulness
H7: OE positively affects Behavioral Intention to Use
H8: Perceived Ease of Use positively affects Behavioral Intention to Use
H9: Perceived Ease of Use positively affects Perceived Usefulness
H10: Perceived Usefulness positively affects Behavioral Intention to Use
H11: Perceived Usefulness positively affects pre-service teacher AI Literacy.
H12: Behavioral Intention to Use positively affects pre-service teacher AI Literacy.
H13: Perceived Usefulness acts as a mediator between Perceived Ease of Use and Behavioral Intention to use.
H14: Behavioural Intention to use acts as a mediator between Outcome Expectancies in Class and pre-service teacher AI Literacy.
H15: Behavioural Intention to use acts as a mediator between Perceived Usefulness and AI Literacy.
H16: Self-efficacy acts as a mediator between Perceived Ease of Use and pre-service teacher AI Literacy.

Research Method

Participants

Non-probability convenience sampling will be employed in this study without a formal sampling frame due to an unknown sampling size before data collection (Teddlie & Yu, 2007). This study will sample from a normal university located in Nanjing, Jiangsu Province, in Eastern China.

To ensure that all participants had been involved in at least one year of study on campus, the first-year students who enrolled for less than one year were excluded from this study. Thus, the targeted participants in the data collection will focus on sophomore, junior, senior and postgraduate-level students in initial teacher education programmes at this university. The estimated sample size is more than 300 pre-service teachers.

Data collection

The data will be collected via an online questionnaire at the end of September 2025. The questionnaire consists of two sections: demographic questions (5 items) and main scales of influential factors (7 scales with 28 items). All the students will participate voluntarily and be informed of the critical information and data processing of this study at a preliminary stage. All the questionnaire responses will be held confidentially and reported anonymously.

Data analysis

The data analysis will involve a two-step procedure, Confirmatory Factor Analysis (CFA) and Structural Equation Modeling (SEM), which were recommended by Hair et al. (2010). CFA will first be conducted on the raw data using IBM SPSS AMOS version 29.0 to examine the construct validity of the measurement instrument. Following that, an SEM using IBM SPSS Statistics version 29.0 will be employed to discover the relationship among the constructs.

Contributions

Theoretical Contribution: By integrating TAM and SCT, this study propose a framework that accounts for both individual cognition and social context influences. It frames AI literacy as a cognitive process shaped by self-efficacy, observational learning, outcome expectations, perceived ease of use, perceived usefulness, and behavioral intention, rather than a single skill or attitude.

Empirical Contribution:Focusing on Chinese pre-service teachers’ AI literacy, this study employs structural equation modeling to examine the applicability of the integrated model in the new cultural context and to explore the interactions among cognitive, social, and motivational factors.

Practical Contribution: This study emphasizes that cultivating AI literacy requires enhancing self-efficacy, creating opportunities for observational learning, and fostering positive outcome expectations. The findings will offer guidance for curriculum, internships, and policy to support pre-service teachers’ development and reflective capacity.

Authors

• YIYUN ZHANG, Univeristy of Brisol

• Yijing Lv, Nanjing Normal University