Introduction

Post-apartheid educational reforms in South Africa recognize the need for integrating technology, including artificial intelligence (AI), to improve the educational experience for teachers and learners (Mokotjo & Mokhele, 2021). However, teachers’ technology adoption, mainly in under-resourced schools, faces multiple barriers, including limited access to affordable devices and the Internet (Chomunorwa & Mugobo, 2023). Some teachers lack adequate technological pedagogical content knowledge (Venketsamy & Hu, 2022), and a general scarcity of resources impedes their ability to integrate technology in teaching and learning (Shilenge & Ramaila, 2020). Structural issues and socio-economic disparities further complicate technology adoption (Mavuru & Ramaila, 2022). Overcoming these challenges necessitates a holistic approach involving resource provision, skill enhancement, curriculum adjustments, and efforts to close the digital divide, enabling more effective technology use in education.

The integration of AI into educational settings has garnered attention in recent years due to its potential to revolutionize pedagogical methods, learner engagement, and assessment practices (Mnguni, 2023). Concurrently, the emergent nature of AI imposes new demands on teachers to adapt and equip themselves with the requisite competencies for successful integration. Within the South African context, there is a paucity of research focusing on the behavioral intentions of teachers regarding AI integration in their teaching. The current research addresses this gap by delineating the behavioral intentions of South African pre-service life sciences teachers for AI integration. Describing these intentions is important for identifying the educational and systemic interventions necessary to foster AI adoption, literacy, and competencies. Moreover, it offers insights into the broader landscape of teacher preparedness and attitude toward emergent technologies. Thus, this research contributes to the existing body of literature and serves as an instrumental resource for policymakers, curriculum designers, and educational stakeholders in optimizing the AI-driven transformation of teaching.

The Emergence of AI in Education

AI is the simulation of human intelligence in machines programmed to perform tasks that typically require human cognitive functions such as learning, reasoning, and problem-solving (Russell & Norvig, 2021). AI is generally categorized into narrow AI, where the system is designed for a specific task, and general AI, which is designed for generalized human cognitive abilities (Sutskever et al., 2014). A notable subclass within general AI is generative AI, which has attracted attention for its capability to produce intricate patterns, texts, or even artworks (Goodfellow et al., 2014).

Understanding these distinctions and subclasses in AI tools is crucial for teachers, as the type of AI integrated into pedagogical settings can impact teaching outcomes (Rizvi et al., 2023). For example, within the education sector, there has been a notable surge in the interest surrounding the application of AI-based instructional tools, AI-enabled learning environments, and AI-supported assessment and feedback systems.

AI-based instructional tools are pedagogical utilities embedded with AI functionalities to augment educational experiences (Chou et al., 2022; Rangel-de Lázaro & Duart, 2023). For example, AI-based instructional tools like natural language processing, machine learning, data analytics, and virtual laboratories can facilitate personalized learning experiences, improving learner engagement and academic performance (Chou et al., 2022). AI-enabled learning environments extend traditional classrooms by cultivating immersive educational experiences (Hashim et al., 2022; Kabudi et al., 2021; Kashive et al., 2020). These settings can vary from virtual to augmented reality platforms, allowing learners to learn through simulated hands-on experiences that foster deeper conceptual understanding and facilitate collaborative problem-solving practices (Cope et al., 2021). AI-supported assessment and feedback systems offer streamlined, efficient evaluative processes that promptly offer personalized, criterion-based feedback, serving formative and summative assessment needs (Hwang et al., 2020; Xu & Ouyang, 2022). They allow learners to benefit from immediate, targeted feedback, thus enhancing learning outcomes (Hwang et al., 2020).

While integrating these AI tools in science education offers promising avenues for advancing pedagogical practices, it is not devoid of challenges that must be addressed. Hence, the teachers’ ability to select and use these tools is integral to fully realizing AI’s transformative potential in science education. Relatedly, their behavioral intentions to employ these tools are vital for effective integration and steering the trajectory of future educational technology deployments.

Teacher Preparedness to Integrate AI into Teaching

While research has shown the potential of AI in enhancing the quality of teaching, the urgency to understand life sciences teachers’ behavioral intentions for integrating AI into their teaching is underscored by a current gap in the literature regarding teacher preparedness for AI adoption in developing countries such as South Africa. Research indicates a deficit in teacher readiness to harness AI innovations (Al Darayseh, 2023; Yue et al., 2024). For example, studies reveal that science teachers frequently have low awareness and understanding of AI applications (AlKanaan, 2022; Shin & Shin, 2020). Furthermore, despite some teachers being cognizant of AI’s potential benefits, a lack of AI readiness creates a disparity between technological advancements and their practical application in educational settings (Chounta et al., 2022; Li et al., 2022).

Researchers suggest that AI-specific competency among teachers could also impact their preparedness to integrate AI into their teaching (Howard et al., 2022; Kohnke et al., 2023). While digital literacy among teachers is increasing, there is a need to cultivate educational AI-specific competencies (Howard et al., 2022; Kohnke et al., 2023). This includes AI literacy as a foundational skill teachers should have (Long & Magerko, 2020; Ng et al., 2021). AI literacy is the understanding and application of AI concepts, ethical considerations, and practical tools (Ng et al., 2021). Research has shown that AI literacy is associated with increased user confidence and preparedness for using AI in practice (Laupichler et al., 2022; Long & Magerko, 2020; Steinbauer et al., 2021). Therefore, AI literacy among teachers could facilitate the effective incorporation of AI tools into teaching practice, thereby enhancing educational outcomes (Al Darayseh, 2023; AlKanaan, 2022; Ayanwale et al., 2022). To this end, teachers need to understand the general characteristics of AI and its pedagogical applications in teaching (Kohnke et al., 2023). They should also possess ethical grounding, technological aptitude, and the ability to meaningfully engage learners through AI (Ayanwale et al., 2022; Karaca et al., 2021).

Given the increasing demands for AI literacy, professional development and training programs targeting these competencies are imperative. Although some efforts have begun to explore teacher preparation through professional development programs, there remains a gap in resources and structured training (Lee & Perret, 2022; Lin & Van Brummelen, 2021). Therefore, ongoing professional development is essential for equipping teachers with the necessary AI skills (Ng et al., 2023). Incorporating AI competencies into pre-service teacher training may also be beneficial to prepare future teachers for the AI-based educational landscape (Ayanwale et al., 2022).

Pre-service Teacher Attitudes Towards AI in Education

Understanding pre-service teachers’ attitudes and behavioral intentions to adopt technologies such as AI for teaching is important as they enter the workforce and will likely face challenges related to implementing AI in their future teaching roles. Unlike in-service teachers, pre-service teachers are at a seminal point in their careers where they are forming the technological pedagogical content knowledge that will underpin their teaching philosophies (Halim et al., 2010; Marshman & Porter, 2013). Their attitudes and behavioral intentions on technology, unencumbered by the inertia of past practices, offer fresh insights into how emerging technologies can be woven into educational contexts.

One of the factors that could affect pre-service teachers’ AI literacy is their exposure and training as students. For example, research shows that pre-service teacher training programs, such as those emphasizing pedagogical knowledge and ICT-related courses, impact their ability to integrate technology into their teaching practices (Aslan & Zhu, 2016). Furthermore, training programs that facilitate the development of technological pedagogical content knowledge can positively impact their readiness to integrate technology into teaching (Akkaya, 2016; Tondeur et al., 2018).

Additionally, pre-service teachers’ attitudes toward technology and perceived ease of use positively influence their competencies for educational practice (Nja et al., 2023; Tondeur et al., 2018). These attitudes may be influenced by perceived competence in exploring emerging technologies for their technology integration in teaching (Min et al., 2023). Attitudes are also nuanced by varying levels of technology readiness, where teachers with low preparedness may feel intimidated by the rapid technological shifts (Chounta et al., 2022). Furthermore, the correlation between self-efficacy beliefs and technology integration tendencies of pre-service teachers emphasizes the significance of self-efficacy in influencing their technology adoption (Caner & Aydin, 2021). It has been suggested that pre-service teachers should be provided with more opportunities to explore teaching with technology to build a positive attitude towards its usage (Wong, 2015).

Pre-service teacher attitudes and behavioral intentions toward technology adoption underscore the urgency of understanding their readiness and intentions as key elements for implementing AI in education (Ayanwale et al., 2022). A focused understanding of their attitudes toward AI could be instrumental in shaping teacher training curricula, policy, and instructional designs, thereby facilitating the adoption of AI in education.

The Current Study

Problem Statement of the Research

While the AI frontier offers remarkable opportunities, it also necessitates a level of preparedness among teachers to employ AI tools in educational settings (Al Darayseh, 2023). In life sciences education, the role of teachers is vital for successfully deploying and utilizing AI technologies (Al Darayseh, 2023; Ayanwale et al., 2022). Within the South African context, a research gap exists, particularly the dearth of studies examining the behavioral intentions of pre-service teachers concerning integrating AI into their future classrooms. Understanding these intentions could be instrumental in tailoring educational interventions, curriculum adaptations, and professional development programs (AlKanaan, 2022). This is because behavioral intentions serve as a predictor for actual behavior, offering a lens through which to anticipate future classroom practices and the extent to which AI will be integrated into life sciences education (Ajzen, 2014; Lauermann & König, 2016). However, the lack of focused research in this area constitutes an urgent research problem. Addressing this issue will provide empirical insights for policy formulation and pedagogical strategy development for teacher training institutions.

Research Aim and Objectives

In light of the above discourse, the current research aimed to qualitatively describe South African pre-service life sciences teachers’ behavioral intentions for integrating AI in teaching. This work is a preliminary effort to understand enablers and barriers to adopting AI in science education. The main research question of the current study is “What are pre-service life sciences teachers’ behavioral intentions for integrating AI in science teaching?”

Theoretical Framework: the Theory of Planned Behavior

The current research utilized the Theory of Planned Behavior as a theoretical framework to address the research aim and question (Fig. 1). Initially posited by Ajzen (2014), the Theory of Planned Behavior provides a social psychological model to predict human behavior based on three constructs, namely, attitude toward the behavior, subjective norm, and perceived behavioral control (PBC) (Ajzen, 2014; Fishbein & Ajzen, 2010). These constructs are further influenced by additional constructs, namely, behavioral beliefs, perceived normative beliefs, and control beliefs.

Fig. 1
figure 1

The Theory of Planned Behavior (adapted from Ajzen, 2014)

Subjective norms and perceived normative beliefs explain external social factors, such as the opinions of mentors or the educational community, which could influence these pre-service teachers’ intentions to integrate AI. Behavioral beliefs consider the perceived outcomes of using AI, where positive beliefs about the benefits can encourage adoption. Perceived subjective norms reflect the social pressures pre-service teachers feel regarding AI use, with support from peers or mentors likely increasing their willingness to integrate AI. Perceived normative beliefs encompass the expectations of significant others about AI usage, which can align closely with pre-service teachers’ intentions. PBC and control beliefs afford the analysis of self-efficacy and the perception of barriers to AI integration, which are significant given the technological and resource constraints often present in South African educational contexts (Ajzen, 2014). Perceived behavioral control assesses pre-service teachers’ perceptions of their capability to integrate AI effectively. Control beliefs identify factors that may facilitate or impede AI use, guiding interventions to overcome barriers. The theory’s unique contribution of PBC addresses behaviors that are not entirely within an individual’s volitional control, as both a direct and indirect influence on intentions and actions (Ajzen & Madden, 1986; Armitage & Conner, 2001). This aspect is relevant for pre-service teachers facing systemic or institutional barriers in employing AI, adding a layer of predictive strength to the Theory of Planned Behavior. As a result, strategies can be developed to enhance the likelihood of successful AI integration in teaching, addressing specific beliefs and norms that influence pre-service teachers’ behavioral intentions.

The Theory of Planned Behavior is not without critics, as some scholars argue that it could benefit from including emotional or moral dimensions (Conner & Armitage, 1998; Sniehotta et al., 2014). However, its proven applicability across various disciplines and its focus on psychological and social variables make it an apt framework for the current research. Empirical studies have demonstrated the effectiveness of the Theory of Planned Behavior in predicting teachers’ technology adoption behaviors, underscoring its relevance for examining AI integration in educational settings (Teo, 2011). By applying the Theory of Planned Behavior, the study methodically examines the psychological and social underpinnings of pre-service teachers’ intentions to integrate AI into life sciences teaching. This framework allows for a nuanced analysis of how each factor—attitudes, subjective norms, and perceived control—contributes to the willingness and readiness of these future teachers to embrace AI technologies (Ajzen, 2014). Such an approach is useful for identifying specific enablers and barriers to AI adoption, offering insights into how educational policies and teacher training programs might be designed to support effective AI integration in science education. The TPB thus provides a robust, theoretically grounded basis for investigating the complex interplay of beliefs, norms, and control perceptions shaping pre-service teachers’ intentions toward AI in education. Therefore, the Theory of Planned Behavior’s comprehensive framework offers a structured approach to assess the multifaceted factors influencing pre-service life sciences teachers’ behavioral intentions, making it highly suitable for the research’s objectives within the South African context.

Methodology

The current research followed an intepretivist research paradigm for its appropriateness in facilitating qualitative research geared towards gathering and interpreting in-depth data to explore and explain phenomena (Creswell, 2014). In alignment with this paradigm, a qualitative interview-based study was adopted using semi-structured interviews for data collection.

Sampling

Purposive sampling was used to recruit participants who met specified inclusion criteria tailored to the research aim. These criteria were as follows: participants needed to be (1) full-time students at a South African university; (2) final-year Bachelor of Education or Postgraduate Certificate in Education students; (3) majored in didactics of life sciences; (4) intent to teach in a government school across varying quintiles; and (5) voluntary participation. Participants were also asked to indicate the quintile classification of the school where they did their most recent teaching practice as final-year students. This was to ensure that selected participants reflect the range of quintile schools in South Africa, where the Department of Basic Education stratifies government schools based on the socio-economic status of their learner population. According to the Department of Basic Education, schools in the first quintile (quintile 1) serve the most economically disadvantaged 20% of learners in a given province. The second quintile (quintile 2) accommodates the subsequent 20% of economically disadvantaged learners, continuing in this pattern until the fifth quintile, which caters to the least economically disadvantaged 20%. Schools within quintiles 1 to 3 are designated no-fee schools, while those in quintiles 4 and 5 operate as fee-paying schools.

The rationale for the selection criteria centered on capturing the experiences and intentions of the participants who are imminently entering the workforce and are likely to face challenges related to implementing AI in their future teaching roles as they had no prior formal training on AI and its use in educational settings. Studying pre-service teachers’ intentions toward AI integration is essential, as they represent the next generation of educators poised to shape the future of teaching and learning. Unlike in-service teachers, pre-service teachers are at a seminal point in their careers where they are forming the pedagogical content knowledge that will underpin their teaching philosophies (Halim et al., 2010; Marshman & Porter, 2013). Their perspectives on AI, unencumbered by the inertia of past practices, offer fresh insights into how emerging technologies can be woven into educational contexts. As potential first-generation AI adopters in South Africa, their behavioral intentions are pivotal, reflecting a blend of innovation readiness and the practicalities of integrating new technologies into pedagogy. Describing these intentions based on empirical research is essential for developing supportive frameworks that align with their developmental needs, ensuring that AI integration enhances educational outcomes in meaningful and sustainable ways.

A total of 10 participants were purposively selected as meeting the pre-set criteria, with two representing each of the quintiles (Table 1). This sample size is within the range of other qualitative studies that have used six to 20 participants to explore constructs guided by the Theory of Planned Behavior (e.g., Huntington et al., 2020; Oppong & Jacob, 2021; Rich et al., 2019; Tan et al., 2015). This sample size also aligns with recommendations for sample sizes in qualitative research, where a minimum of five participants is considered suitable (Hennink & Kaiser, 2022). Furthermore, the selected sample was deemed suitable since qualitative research focuses on obtaining deep, detailed insights into participants’ experiences, perceptions, and motivations, emphasizing the quality of data over quantity (Aspers & Corte, 2019; Robinson, 2014). It employs an inductive approach, where understanding the complexity of human behavior in specific contexts is more valuable than generalizing findings to a larger population (Creswell, 2014). This approach necessitates a smaller, more manageable sample size, allowing researchers to conduct thorough analyses and engage deeply with each participant’s responses, thereby uncovering rich, nuanced understandings of the phenomena under study. The sample size was further justified by referring to the principle of informational redundancy, whereby additional probing is unlikely to provide new insights after a certain point (Malterud et al., 2016). This sample size also afforded a rich, context-specific understanding of the participants’ perspectives, enhancing the research’s internal validity (Creswell & Poth, 2018).

Table 1 A sample description of the research’s participants

Data Collection

Data were collected through individual online interviews utilizing an interview protocol with open-ended items designed to qualitatively describe pre-service life sciences teachers’ behavioral intentions toward integrating AI in science teaching. To address each research objective, the interview protocol comprised six items, each corresponding to the construct of the Theory of Planned Behavior construct as described by Ajzen (2006) and Fishbein and Ajzen (2010). To this end, objective 1, which probed attitudes towards integrating AI, was contextualized by asking respondents about their perceptions of the usefulness and appropriateness of incorporating AI in their life sciences curriculum. This sought to unearth underlying attitudes that might facilitate or hinder AI adoption. Objective 2 was framed around eliciting behavioral beliefs by querying respondents on AI implementation’s perceived advantages and disadvantages in life sciences teaching. This provided a landscape of teachers’ expectations and concerns. Objective 3 aimed to gauge subjective norms by asking respondents to discuss the level of support or discouragement they perceive from colleagues and administrators regarding AI integration. This gave insight into the social pressures influencing teachers’ decisions. Objective 4 explored normative beliefs by identifying the individuals or groups that most influence respondents’ decisions on AI adoption. Objective 5 examined perceived behavioral control, focusing on respondents’ self-efficacy in incorporating AI. The interview item here probed the factors contributing to their confidence or lack thereof. Objective 6 explored control beliefs by asking about the resources, training, or support respondents believe are necessary to overcome AI integration barriers. These objectives provided a multifaceted understanding of the complex variables affecting AI adoption in life sciences education.

Using open-ended questions allowed the researcher to probe the participants’ responses until they yield no further theoretical insights or new properties (Hennink & Kaiser, 2022). This enabled the researcher to delve deeply into participants’ perspectives without constraining their responses to predefined options. This flexibility allowed for the exploration of nuanced and detailed insights. By probing further with follow-up questions, the researcher exhaustively explored each aspect until no new information emerged, thus reaching saturation (Saunders et al., 2018). This iterative process ensured a comprehensive understanding of participants’ perspectives, capturing the richness and diversity of their experiences and ensuring that the analysis was grounded in their own words and contexts.

The interview protocol was first validated through a panel of experts who determined face, construct, and content validity to ensure the instrument was fit for purpose (Flouris et al., 2010). In particular, similar to Martínez and Llauradó (2020), the panel determined the extent to which the protocol measures the theoretical construct it is intended to measure. Having satisfied the panel of experts, the instrument was piloted on respondents who resembled the primary sample’s characteristics. This was to confirm the face validity of the instrument. Improvements recommended by the panel of experts and the pilot group were used to improve the instrument. The interviews range between 31 and 47 min, generating between 2204 and 3854 words of data per interview (Table 1).

Data Analysis

In this study, Braun and Clarke’s (2006) six-step process for thematic analysis was employed to generate themes and sub-themes. Initially, the interview transcripts were read multiple times to familiarize the researcher with the data. The second step involved generating initial codes line-by-line, focusing on the explicit and implicit meaning of text segments related to the integration of AI (Table 2). Step 3 aggregated these codes into potential sub-themes based on recurring patterns. In the fourth step, overarching themes were defined to encapsulate the sub-themes conceptually. The fifth step involved refining themes, ensuring they coherently represented the coded data. Lastly, themes were finalized after a thorough review, providing a nuanced understanding of pre-service teachers’ perspectives on AI integration in life sciences teaching.

Table 2 The analysis of interviewee responses when asked about their perceptions of the usefulness and appropriateness of incorporating AI in their life sciences curriculum

To ensure the trustworthiness and credibility of the analysis, an additional independent researcher with expertise in qualitative research methods independently reviewed the coded data and the themes identified. This was to enhance the dependability and confirmability of the data analysis. Specifically, the independent researcher independently cross-verified and reviewed the primary researcher’s analysis, highlighting missed elements and challenging interpretations. This led to a more nuanced and critical analysis and interpretation of the data (Lincoln & Guba, 1985; Mnguni, 2024). This process refined and bolstered the arguments and conclusions derived from the data. Data were further scrutinized in instances of differing opinions to resolve these conflicts. Consequently, the conclusions formulated were grounded on the available data.

Results

Summary of Themes Emerging from the Interviews

The thematic data analysis revealed several themes (Fig. 2) within each construct of the Theory of Planned Behavior that framed the interviews. For example, under attitudes toward AI integration in the life sciences curriculum, themes identified included pedagogical benefits, practical limitations, and philosophical concerns. Themes related to behavioral beliefs were the advantages and disadvantages of AI integration. Concerning subjective norms, the emerging themes were inter-generational differences, administrative issues, community roles, technology perceptions, and resource constraints. The perceived normative beliefs had organizational authority, peer influence, parental concerns, and policy and funding as the emerging themes. Emerging themes are presented in Tables 3, 4, 5, 6, 7, and 8, with corresponding sub-themes and evidence.

Fig. 2
figure 2

The emerging themes within each of the constructs of the Theory of Planned Behavior concern pre-service teachers’ behavioral intentions for integrating AI into life sciences teaching

Table 3 Attitudes towards the integration of AI technologies life sciences curriculum
Table 4 Behavioral beliefs about the advantages or disadvantages of integrating AI in life sciences teaching practices
Table 5 Perceived subjective norms about the integration of AI in life sciences teaching
Table 6 The perceived normative beliefs about the integration of AI in life sciences teaching
Table 7 The perceived behavioral control about the integration of AI in life sciences teaching
Table 8 The control beliefs about integrating AI into life sciences teaching

Themes, Sub-themes, and Evidence Concerning the Integration of AI in Life Sciences Teaching

Results (Table 3) demonstrated varied attitudes among pre-service life sciences teachers regarding AI integration. The themes that emerged highlighted both the potential pedagogical benefits and challenges. For instance, the participants highlighted AI’s capacity to make learning more interactive, personalize education, and render complex topics like DNA sequencing more accessible as a benefit. However, there were concerns about practical limitations in resource-constrained settings. Teachers from rural areas underscored the challenges of limited technological infrastructure and accessibility issues. Philosophical concerns also emerged, emphasizing potential dependence on technology, a possible undermining of traditional teaching, and the importance of maintaining cultural and ethical considerations within the curriculum. The themes highlighted a balance between embracing AI’s benefits and being cognizant of its limitations.

Regarding teachers’ behavioral beliefs about integrating AI in life sciences teaching practices, results delineated both advantages and disadvantages (Table 4). The participants identified advantages, including enhancing learner engagement, personalized learning experiences, increased efficiency, and promoting interactive learning. The participants underscored the potential for AI to transform traditional pedagogy by individualizing feedback, automating tasks, and fostering active learner participation. However, concerns were also reported as disadvantages, which include issues of accessibility, the digital divide, the steep learning curve for teachers, and the time needed for teacher training. Potential costs associated with integrating AI in resource-limited schools were also highlighted. These results suggest that while pre-service teachers are aware of the potential benefits of integrating AI into their teaching, they are not oblivious to the challenges facing the education system, which may impede the successful integration of AI.

Data also revealed several themes concerning perceived subjective norms about integrating AI in life sciences teaching (Table 5). For example, participants reported perceived inter-generational differences. Specifically, participants suggested that there are differences in how younger and older teachers perceive the integration of AI. Younger teachers are perceived as supportive, while older teachers are perceived as skeptical. Participants also perceived administrative perspectives as results-driven and limited by budget constraints. They were perceived as potentially skeptical, prioritizing basic educational needs over technological advancements. Technology perceptions among the participants varied, with some participants recognizing its benefits while others expressing concerns about the potential adverse impact on critical thinking. The participants emphasized resource constraints, pointing to gaps between intentions and reality, and highlighted the need for proper teacher training.

Data analysis related to the perceived normative beliefs suggested that organizational authority could influence the integration of AI into life sciences teaching (Table 6). To this end, participants highlighted the role of the School Governing Body and the control of resources and policy. Some of the participants highlighted the dominant role of the School Governing Bodies in decision-making and resource allocation. Peer influence underscored the role of fellow teachers, including experienced teachers and heads of departments, as influential figures. Perceived parental concerns among participants were evident as they indicated that parents’ skepticism about AI’s role in education could hinder its integration. The policy and funding theme demonstrated that participants perceive the Department of Basic Education’s paramount role regarding budget and curriculum decisions.

Results indicated that pre-service life sciences teachers had varied perceptions regarding their perceived behavioral control for integrating AI in teaching (Table 7). One of the themes was confidence levels, where some participants reported low confidence, while others were highly confident. Moderate to high confidence was rooted in familiarity with technology. Training and resource availability emerged vital, with participants reporting that they needed AI training. Some participants raised concerns about the compatibility of AI with their teaching styles, mainly for teachers who preferred hands-on interactions. Furthermore, apprehensions regarding AI’s impact on learners were voiced, with concerns about equity and potential cognitive implications. These results underscore the multifaceted factors influencing teachers’ behavioral intentions toward AI integration.

Results concerning control beliefs about integrating AI into life sciences teaching revealed several themes, including training and education, technical support, resource availability, policy, and funding (Table 8). Regarding training and education, participants suggested that they require hands-on workshops and specialized and comprehensive training. Some participants expressed the need for continuous mentoring and support rather than one-off sessions. Technical support emphasized the need to have IT staff available to assist where teachers may lack technical expertise. Resource availability highlighted the necessity of reliable hardware, internet, and electricity for successful AI integration. Under policy and funding, participants stressed the need for institutional support, funding for AI tools, and evidence-backed decisions for AI incorporation in teaching.

Discussion

The primary finding of this research reveals a complex interplay of factors influencing pre-teachers’ behavioral intentions as determined through the constructs of the Theory of Planned Behavior. This finding echoes research that shows multiple factors affecting teachers’ intentions to use technology in their teaching (e.g., Teo, 2011).

This research has found mixed attitudinal perspectives, with participants appreciating AI’s pedagogical benefits but wary of practical and philosophical limitations. Behavioral beliefs outlined advantages, such as enhanced learner engagement and efficiency, and disadvantages, including the digital divide and resource constraints. Subjective norms indicated that younger teachers are perceived to support AI integration, whereas older teachers and administrative bodies are believed to be more skeptical, often due to budget constraints. Perceived normative beliefs highlighted the role of organizational authority and peer influence in decision-making, with parents’ skepticism identified as a potential hindrance. Concerning perceived behavioral control, teachers’ confidence levels varied and were often tied to technological fluency. Control beliefs emphasized the need for comprehensive training to enhance AI literacy among teachers, as well as technical support and robust policy measures. These findings suggest that while teachers value integrating AI, multiple layers of beliefs, norms, and resource availability shape their behavioral intentions toward actual implementation.

The current research findings generally echo existing literature, even though some contradict the literature. For example, the findings on the role of organizational authority in AI integration in life sciences education corroborate existing research emphasizing the influence of institutional governance in educational technology adoption (Arief, 2021). Additionally, the importance of peer influence aligns with social constructivist theories, suggesting that peer interactions impact teachers’ acceptance of technological change (Vygotsky, 2018). However, the current research departs from the literature on perceived parental concerns. While previous research suggests that parental skepticism is generally low regarding educational technology (Osorio-Saez et al., 2021), current findings indicate a palpable concern among parents about AI’s role, serving as a potential hindrance. In terms of perceived behavioral control, our research is consistent with the self-efficacy theory and recent research which shows that confidence levels are a determinant factor in adopting new technologies (Bandura, 2002; Park et al., 2023). However, the emphasis on training and resource availability in our research is more acute than in prior studies, considering the specialized nature of AI technologies. Moreover, the explicit concerns about AI’s impact on learners found in this research seem less prominent in existing literature, thus opening a new avenue for research into the pedagogical implications of AI in education.

The research also found that the quintile classification of schools where pre-service teachers conducted their teaching practice shapes their perceptions and readiness toward integrating AI in life sciences teaching. This finding supports existing literature that shows that work-integrated learning, including teaching practice, shapes teachers’ pedagogical content knowledge (e.g., Myers & Gray, 2017; Rusznyak & Bertram, 2021). For example, teachers exposed to schools predominantly catering to economically disadvantaged learners exhibit greater concerns about resource limitations, emphasizing practical constraints and philosophical considerations regarding cultural and ethical aspects of AI integration. In contrast, those from well-resourced schools express concerns about policy and funding implications. They emphasize policy influence and funding availability in their behavioral and perceived normative beliefs, indicative of different challenges and priorities. This disparity highlights how the economic context of schools shapes pre-service teachers’ perceptions and concerns about integrating AI into their teaching practices. While not widely reported, this phenomenon is related to previous findings that show that in-service training impacts teachers’ understanding and acceptance of educational technology (Gumbo et al., 2012). Given that teaching practice is a form of work-based learning, these findings have implications for student placement during their teaching practice. Institutions of teacher education should be circumspect about environments that pre-service teachers are exposed to during pre-service training, as this could impact their long-term behavioral intentions.

Implications of the Findings of the Research

The findings highlight several implications for policy and practice in integrating AI into life sciences teaching. Firstly, the role of organizational authorities, such as School Governing Bodies and the Department of Basic Education, calls for targeted policies to address resource allocation to support AI integration and teacher support and development. Since these bodies control budgets and curricula, their commitment is essential for funding AI initiatives and updating curricula to incorporate AI-enabled pedagogies. Secondly, the variance in perceived behavioral control among teachers underscores the need for comprehensive AI literacy training programs. Training should improve technical skills and focus on pedagogical approaches that align with new technologies. The necessity for ongoing mentoring and technical support further reveals that sustainable professional development model is more effective over episodic training sessions. Lastly, concerns about teaching styles and learners’ abilities indicate that AI integration must be flexible to accommodate diverse pedagogical approaches and learner needs (Ertmer & Ottenbreit-Leftwich, 2013). This calls for the creation of adaptable AI tools that can be customized to different classroom settings, learner proficiencies, and policies that ensure equitable access to technology.

Limitations of the Study and Recommendations for Future Research

While providing valuable insights into the factors affecting the integration of AI in life sciences education, studying pre-service teachers limits this research as these participants have yet to fully engage in the practicalities of teaching, lacking the nuanced understanding that comes with classroom experience. Their intentions toward AI integration might only partially capture the complexities of teaching contexts. Future research should explore in-service teachers’ perspectives to understand how teaching experiences influence AI adoption and describe the challenges and opportunities that seasoned educators face in integrating AI into their pedagogical practices.

Given the qualitative nature of this research, its limitations include the inability to generalize findings broadly across different educational contexts. Future research should employ mixed methods or large-scale quantitative approaches to substantiate its findings. There is also a need for in-depth studies focusing on the specialized AI literacy training requirements for AI in life sciences and exploring the ethical and pedagogical implications of AI use in classrooms (Ertmer, 2005).

Recommendations for future work include investigating the efficacy of different AI-enabled pedagogical approaches and developing policies that ensure equitable access to AI technologies across diverse educational settings. Finally, longitudinal studies can help explain the long-term effects of AI integration on learner outcomes and teacher practice.

Conclusion

Based on the findings of the current research, it is concluded that the behavioral intention of pre-service life sciences teachers to integrate AI in life sciences teaching is moderate to low for several reasons. Attitudes towards AI are mixed, while there is optimism about AI’s potential, reservations concerning efficacy and learner engagement exist. Subjective norms underscore that integration is contingent upon multiple stakeholders, such as School Governing Bodies, senior teachers, and parents, who present additional barriers and expectations. Perceived behavioral control shows a diverse landscape. Concerns over inadequate training, resource availability, and compatibility with teaching styles indicate that even if teachers were willing, they may lack the required skills or resources to integrate AI. Therefore, despite some favorable conditions, the confluence of these TPB constructs suggests a moderate to low behavioral intention of pre-service life sciences teachers to integrate.