Drivers of Chatbot Adoption among K–12 Teachers in Saudi Arabia

Abstract

The aim of the current study was to identify the factors influencing teachers’ acceptance of using chatbots in education. To achieve this, we employed the descriptive method, applying the conceptual framework of the unified theory of acceptance and use of technology (UTAUT2) to a sample of 406 teachers from the Eastern Province in Saudi Arabia. By applying structural equation modelling (SEM), the research results revealed that the most influential factors of chatbot acceptance among Saudi teachers were artificial intelligence literacy, followed by hedonic motivation, and then social influence. Meanwhile, performance expectancy, effort expectancy, and facilitating conditions were not significant indicators of teachers’ acceptance of using chatbots in education. In light of these findings, we recommend the following actions: focusing on artificial intelligence literacy for teachers, holding specialised workshops on chatbot applications in education, enhancing teachers’ motivation to use chatbots, and forming professional communities for teachers to exchange experiences and knowledge.

1. Introduction

The twenty-first century has seen rapid advancements in various fields, especially information and communications technology. This surge requires individuals and institutions to adapt and harness technological innovations to maximise benefits. While machines have traditionally lacked intelligence and have typically been associated with routine tasks, the growing interest in artificial intelligence (AI) has resulted in substantial advancements in comprehending and replicating human intelligence within machines. These advancements have transformed human–machine interaction, with AI applications emerging as key innovations [1].

AI systems are set to make substantial contributions to education, particularly in developing learning environments. Among the emerging technologies are chatbots, which have become digital assistants. These chatbots engage in intelligent dialogue and enhance learning environments by providing performance support and delivering learning content on demand, thus offering a more flexible and effective educational experience [2]. Chatbots are demonstrating increasing proficiency in executing a broad spectrum of tasks, encompassing both routine operations and more complex, sophisticated activities [3,4].

Chatbots exemplify generative AI, as they use deep neural networks to create new data [5]. They are one of the most prominent forms of advanced human–machine communication [6,7]. Defined as conversational programs that interact with users through natural language, chatbots automate conversations by simulating human dialogue and are integrated into platforms and messaging services [8,9]. Their ability to engage a broad audience effectively positions them as tools for information gathering, blurring the line between human and machine interactions [10].

Successfully integrating new technology into education requires an understanding of educators’ reasons for accepting or rejecting it. The success of interactive chatbots in educational settings depends largely on teachers’ perceptions of their effectiveness [11]. Analysing these perceptions is crucial, as teachers are key stakeholders in the educational process and are essential to the successful adoption of technological innovations [12]. This study adopts the unified theory of acceptance and use of technology (UTAUT2), as introduced by Venkatesh et al. [13], with an extension that includes AI literacy (AIL). The aim is to achieve a deeper insight into teachers’ competence in using AI applications and its relevance to the study topic, considering this to be a new element that, to the best of our knowledge, has not been measured with the UTAUT2 before. AIL involves the ability to understand and apply AI technology, critically evaluate data, and promote a sense of accountability and respect for mutual rights and responsibilities [14].

2. Aims and Significance

Saudi Arabia’s Vision 2030 emphasises the importance of advancing the education sector by adopting a comprehensive plan to enhance the educational system through the implementation of global best practices [15]. AI is closely tied to the future of education, and its continued development is crucial for achieving the Kingdom’s long-term goals. However, the current application of AI in Saudi Arabia’s educational sector remains in its early stages, compared to other countries [1]. This necessitates further research by experts to ensure the optimal use of AI in achieving the goals of Vision 2030.

The focus on educational chatbots is rising, and we recognise them as a key development for a technologically enhanced educational future [16,17,18]. Chatbots simulate human conversations, offering a more natural interaction method and providing a unique educational experience by mimicking human dialogue. This helps address the lack of flexibility in learning systems and enhances students’ engagement with the system [19]. Teachers play a crucial role in creating interactive learning environments and making chatbots valuable educational tools in this context [20].

Teachers’ perceptions of chatbots significantly influence how they deliver content, engage with technology, and integrate it into their teaching. If teachers perceive that chatbots contribute to student learning, they are more inclined to integrate them into their instructional methods [21]. Conversely, if they do not see the benefit, they may avoid using them, highlighting how critical teachers’ beliefs are to the usefulness of educational chatbots [11].

Most experimental studies have focused on the effectiveness of chatbots in general education [22,23,24] and higher education [25,26]. However, few studies have examined the specific intentions of K–12 teachers to use chatbots in education [11]. Research examining the factors that influence teachers’ intentions to use chatbots in K–12 education, especially in Saudi Arabia, is significantly limited. In response to this gap, there is a need for studies that offer a thorough exploration of the determinants that shape K–12 teachers’ adoption of chatbots in the educational process [27,28].

The importance of utilising AI tools in modern education has been emphasised. Previous literature has also recommended new research to investigate the factors influencing teachers’ acceptance of chatbots in education [12], expand the application of the UTAUT2 by testing new variables and other educational technologies [29], and understand the factors that affect a specific sample of teachers [30], particularly considering the scarcity of studies addressing AIL in the educational field [31]. In light of these, this study seeks to investigate the factors that influence the acceptance of chatbots among K–12 teachers in Saudi Arabia, and to understand the influence of AIL. Understanding teachers’ perspectives on chatbots and their intentions to use them could reveal new directions for developing educational chatbots [20].

3. Chatbots in Education

Chatbots are software programs that typically operate within instant messaging platforms, such as Facebook Messenger and Telegram. They provide users with instant assistance around the clock, using predefined scenarios without the need for human intervention. Przegalinska et al. [32] describe chatbots as electronic programs that facilitate conversation through artificial intelligence entities (digital entities). These chatbots have interfaces that interact naturally with users to conduct text or voice conversations, analyse incoming messages, and respond in various ways depending on who is speaking, what they are saying, and the topic of conversation. Users are increasingly engaging with these digital entities in a manner similar to their interactions with one another. Chatbot technology has risen in prominence within the digital landscape, owing to its capability to respond 24/7 and handle multiple requests simultaneously—something that would be impossible for a single person to manage [33].

Chatbots are becoming increasingly recognised as important tools in both administrative and educational settings, leveraging artificial intelligence to support personalised learning and tasks that require repetitive practice [10,34,35]. These chatbots not only answer common questions but also assist teachers by reducing their workload, allowing them to focus on research and improving their teaching methods [5]. Today, education requires tools that are easy to set up and can be used in various learning environments. Chatbots fulfil this need by providing instant responses to user inquiries, offering accurate information on specific topics and carrying out tasks efficiently [36]. Teachers can use these chatbots to monitor student engagement and understand learning behaviours by analysing student interactions. This technology helps teachers manage their classrooms more effectively and shifts their role towards facilitating learning [37].

Kasneci et al. [38] note that chatbots are valuable for lesson planning. They enable educators to create curricula and design a variety of practice exercises that encourage the participation of students with different learning abilities. These chatbots also support language learning by providing text translations, suggesting grammar improvements, and simulating conversations to help students improve their communication skills. This process builds student confidence and fosters active engagement in learning. Moreover, chatbots offer a flexible learning environment and tailored educational content, allowing students to explore topics of interest more deeply [39]. These tools recommend materials that match individual students’ needs and allow quick access to content at any time and place, making the learning process more focused on achieving specific outcomes [33,40,41].

Given the influence of these chatbots on education, it is essential that teachers adapt their skills and teaching practices to meet these new demands. Teachers play a vital role in decision-making, and their acceptance of and intentions to use chatbots must be understood, as they are the key actors in integrating this technology into the educational process [11]. Therefore, the successful implementation of chatbots requires positive attitudes from teachers, which can present an additional challenge when utilising chatbots in the classroom [42].

4. Acceptance of Chatbot Technology in Education

The integration of chatbots in educational settings has garnered increasing interest, leading to a variety of studies investigating factors that influence their acceptance. The Technology Acceptance Model (TAM), the UTAUT, and its extended version UTAUT2 have been widely used to understand the dynamics behind technology adoption in education, particularly concerning chatbot utilisation among students, teachers, and pre-service teachers.

An et al. [43] utilised the UTAUT to model English teachers’ behavioural intentions to use AI, including chatbots, in middle schools. The study identified performance expectancy, facilitating conditions, and social influence as key factors influencing teachers’ intentions. This research underscores the need for support mechanisms, such as professional development and institutional encouragement, to enhance AI’s adoption in educational settings.

Al-Abdullatif [34] integrated the TAM with the value-based adoption model (VAM) to model students’ perceptions of chatbots in learning. This approach not only examined the perceived ease of use and usefulness, as proposed by the TAM, but also considered perceived value, satisfaction, and other factors that might affect students’ acceptance and use of chatbots. The findings suggest that when chatbots provide clear learning value and enhance students’ satisfaction, their adoption in educational settings is more likely.

Chocarro et al. [12] investigated teachers’ attitudes towards chatbots in educational settings using the TAM. The results indicated that teachers’ positive attitudes towards chatbots were significantly influenced by the proactivity of the bots, particularly in enhancing educational interactions. The study emphasised that chatbot design should prioritise social language features to foster proactive engagement, thereby improving teachers’ perceptions of their utility in educational contexts.

Al Darayseh [44] examined science teachers’ acceptance of artificial intelligence, including chatbots, using a TAM-based approach. The study found that science teachers’ attitudes towards AI and perceptions of its usefulness significantly influenced their acceptance. The research also indicated that teachers were more likely to accept AI technologies if they believed that these tools could enhance their teaching practices and support students’ learning outcomes. The findings highlight the importance of demonstrating the practical benefits of AI in education to encourage teachers’ adoption and integration of AI into their pedagogical practices.

Yang and Chen [11] explored pre-service teachers’ perceptions and intentions regarding chatbot use through statistical and lag-sequential analysis. The study found that pre-service teachers exhibited a generally positive perception of chatbots, influenced by factors such as ease of use, usefulness, and potential to support teaching practices. The findings underscore the importance of integrating chatbot training into teacher education programmes to enhance pre-service teachers’ acceptance and effective use of chatbots in future classrooms.

Strzelecki et al. [45] examined the acceptance and use of chatbots within the academic community, employing a UTAUT-based framework. The study indicated that perceived ease of use, usefulness, and social influence were crucial in shaping the attitudes and behavioural intentions of academic users towards chatbots. The authors argued that incorporating chatbots in academic settings requires a focus on enhancing their usability and demonstrating their value in supporting academic tasks. This research points to the potential of chatbots as supportive tools in educational environments, emphasising the need for further exploration of their integration strategies.

Li et al. [46] extended the classic TAM model to explore factors influencing engagement with AI in education at the K–12 level. The study identified that perceived usefulness, ease of use, and subjective norms significantly impacted students’ engagement with AI, including chatbot technologies. Furthermore, the research highlighted the importance of integrating AI education into K–12 curricula, suggesting that early exposure to AI tools, such as chatbots, can positively shape students’ attitudes and acceptance, ultimately leading to more effective adoption in educational contexts.

Yang and Appleget [47] applied the TAM to investigate pre-service teachers’ perceptions of generative AI, including chatbots. The study revealed that perceived usefulness and perceived ease of use significantly affected pre-service teachers’ intention to use generative AI in their teaching practices. The authors suggested that addressing concerns around the complexity of AI technologies and demonstrating their practical utility in teaching could promote more positive attitudes and increase the likelihood of adoption among future educators.

5. Theoretical Framework and Literature Review

Chatbots present a new model for empowering both learners and educators by enhancing the educational process across various stages and domains. Effective utilisation of these chatbots can significantly improve educational outcomes and support professional development. As such, understanding teachers’ perceptions of accepting and integrating this technology is crucial. Teachers’ acceptance is a key indicator of the successful implementation of educational technology, and studying the behavioural factors that influence their acceptance has become a focal point of research [48]. Various theories have been formulated to analyse how individuals adopt new technologies, with the UTAUT standing out as one of the most influential. Introduced by Venkatesh et al. in 2003, the UTAUT offers a robust framework for examining how individuals accept and utilise technology. The original model highlights several critical factors that shape technology’s adaption: performance expectancy, effort expectancy, social influence, and facilitating conditions. In 2012, Venkatesh and colleagues expanded this model into UTAUT2 by adding elements such as hedonic motivation, price value, and habit. These additions enhanced the model’s explanatory power, accounting for 74% of the variance in behavioural intention to use technology, thus making UTAUT2 particularly valuable during the preliminary stages of technology adoption [49].

In the current study, the UTAUT2 model was adapted to the context of chatbots, focusing specifically on teachers’ intentions to use chatbots (IUChatbot). The model was modified to include five core factors: performance expectancy (PEX), effort expectancy (EEX), social influence (SIN), facilitating conditions (FCOs), and hedonic motivation (HMO). Additionally, AIL was introduced as a sixth factor. The inclusion of this factor aimed to assess teachers’ awareness and understanding of AI applications and tools, as well as teachers’ ability to evaluate their ethical use. This addition is supported by recommendations from several studies that suggest examining the role of AIL in predicting the acceptance of educational chatbots [14,34,50]. AIL is also an intriguing new variable to explore with the proposed research in the educational context (Figure 1).

The original UTAUT2 factors—specifically, price value, habit, and actual use—were omitted from the model due to their unsuitability to this study’s context. Price value was considered irrelevant since educational institutions often provide technology resources, such as chatbots, at no additional cost to teachers. Since teachers are not typically responsible for purchasing or subscribing to these tools, price value may have little to no relevance in their decision-making processes. In terms of habit, this factor was removed because it describes the automatic use of technology resulting from familiarity. However, most teacher participants reported not having prior experience with interactive chatbots in the educational process. Likewise, our study aimed to explore teachers’ perceptions and acceptance of chatbots, focusing on the factors that could influence their willingness to adopt this technology in the future. Including the actual use would not have been feasible in this specific context, as most of the participants have not yet had sufficient experience to accurately report on this aspect. The following section provides a literature review of the factors of the proposed model.

5.1. Performance Expectancy (PEX)

PEX is frequently cited as a crucial determinant of an individual’s intention to adopt technology [16]. Following Venkatesh et al.’s [51] definition, PEX refers to the extent to which using a particular technology is perceived to enhance the efficiency and effectiveness of completing a specific task. In the context of the present study, PEX pertains to how strongly teachers believe that employing chatbots will assist them in fulfilling their educational responsibilities more effectively and efficiently. Regardless of how user-friendly a technology may be, its adoption is improbable if it is not seen as valuable. Therefore, the likelihood of users adopting chatbots increases when they perceive these tools as beneficial to them [52]. For example, if teachers perceive that interactive chatbots can reduce their workload, improve student engagement, or streamline communication, they will be more inclined to integrate these technologies into their classrooms. In the educational context, numerous studies have identified PEX as a major factor that influences users’ intention to adopt digital technology [27,52,53,54,55]. As a result, we propose Hypothesis 1 (H1): PEX is a determinant of teachers’ IUChatbot.

5.2. Effort Expectancy (EEX)

EEX plays a crucial role in the acceptance of technology [56]. It is described as the level of ease associated with using a particular system [13]. In the context of education, EEX holds significant importance, as teachers’ willingness to adopt technology often hinges on how effortlessly they can incorporate it into their teaching and learning activities. For instance, when teachers find a chatbot’s interface intuitive and easy to navigate, their likelihood of adopting the technology increases [57]. Alternatively, if a chatbot is seen as cumbersome or requires extensive effort to learn and implement, it may face resistance, even if it offers substantial educational advantages. Numerous educational studies have emphasised that EEX is a key determinant that affects users’ intention to accept new digital technologies [53,54,55]. Therefore, we put forward Hypothesis 2 (H2): EEX is a determinant of teachers’ IUChatbot.

5.3. Social Influence (SIN)

SIN refers to the extent to which individuals believe that significant people in their work domain think that they should use a certain technology [13]. In the context of education, this implies that teachers are more likely to integrate digital technologies into their practices if they perceive that influential figures—such as school administrators, students, fellow teachers, and other key stakeholders—are supportive of their adoption [58,59]. In the current study, we focus on whether teachers were inclined to use chatbots with their students if they sensed that influential figures, such as fellow teachers, colleagues, parents, and school leadership, endorsed their use. SIN has been acknowledged in numerous studies as a significant predictor in determining whether individuals intend to adopt a particular technology [53,59,60]. Drawing from this, we propose Hypothesis 3 (H3): SIN is a determinant of teachers’ IUChatbot.

5.4. Facilitating Conditions (FCOs)

Venkatesh et al. [13] identified FCOs as a significant determinant of technology use intention. This determinant is closely tied to users’ perceptions of the support and resources available for task performance. These conditions encompass both technical and organisational infrastructures that support system usage, such as the provision of access to essential knowledge and skills, as well as an environment that promotes and supports individuals’ willingness to adopt technology [53]. In this study, FCOs pertain to teachers’ views on the preparedness of the technological infrastructure and the level of institutional support available for the integration of chatbots into the educational process, along with their confidence in possessing the required knowledge to effectively leverage this infrastructure. Previous research has demonstrated that FCOs play a crucial role in influencing behavioural intentions [53,59]. This suggests that when teachers perceive adequate support and resources, they are more inclined to utilise new technologies, such as chatbots, in their teaching practices. Accordingly, we put forward Hypothesis 4 (H4): FCOs are a determinant of teachers’ IUChatbot.

5.5. Hedonic Motivation (HMO)

HMO reflects the degree of pleasure and enjoyment that users derive from engaging with a technology [51]. In educational settings, this can translate to the extent to which teachers enjoy using digital learning tools, which can significantly impact teachers’ willingness to integrate these tools into their classrooms. In this study, we focus on the enjoyment teachers perceive from using interactive chatbots in their classrooms. This enjoyment can stem from various factors, such as the ease of interaction with the chatbot, the novelty of the technology, or the enhanced engagement that it facilitates among students. When teachers find the use of chatbots enjoyable, they are more inclined to adopt them, which can lead to more innovative and dynamic teaching methods. Several studies have established a strong connection between motivation and individuals’ intentions to adopt digital technology [27,57,58,61]. Considering this, we propose Hypothesis 5 (H5): HMO is a determinant of teachers’ IUChatbot.

5.6. AI Literacy (AIL)

Given current technological advancements and future projections, job roles will increasingly be tied to AI, which will underpin the products upon which our future wealth is built. In this context, AIL will become as crucial as traditional literacy (reading and writing) and a key issue in the future [62]. AIL refers to an individual’s capability to comprehend essential AI concepts and competencies. AIL is further described as encompassing a range of skills that enable effective and critical interaction with AI applications [63]. Moreover, Ng et al. [31] emphasise the importance of including AIL in the broader spectrum of digital literacies necessary for living in the 21st century. They recommend that educators update their AI knowledge and leverage available technologies to develop AIL. This encompasses the utilisation of AI-driven tools, such as adaptive learning platforms, chatbots, and translation applications, which facilitate teaching practices and enhance personalised learning by understanding student progress and needs, ultimately improving the learning process and overcoming teaching challenges.

Digital literacy research has consistently demonstrated close ties between users’ digital literacy and their adoption of new technologies [14,64,65]. Therefore, it is reasonable to conclude that AIL will similarly be positively associated with individuals’ attitudes towards AI and the daily use of its applications and products. Chai et al. [66] found that students’ intentions to use AI applications could be determined by examining their AI literacy, aligning with previous research on the factors that influence technology use. On account of this, we formulate Hypothesis 6 (H6): AIL is a determinant of teachers’ IUChatbot.

6. Methods

6.1. Sample Demographics and Data Collection Process

In March and April 2023, data were collected through an online survey distributed to a broad sample of teachers across the country. The survey was administered after obtaining ethical approval from the university’s review board, ensuring that all participation was voluntary, confidential, and used exclusively for academic research purposes. The sample comprised 406 educators, predominantly female (78.08%), with males representing 21.92% of the participants. The age distribution varied, with 22.17% of respondents aged 20–29, 32.02% aged 30–39, 32.76% aged 40–49, and 13.05% aged 50 and above. Teaching experience among the participants was also diverse: 28.82% had less than 5 years of experience, 23.40% had between 5 and 10 years, 25.12% had between 11 and 20 years, and 22.66% had over 20 years of experience. The educators were involved in teaching at different educational levels: 28.57% at the elementary level, 21.43% at the middle school level, 28.08% at the high school level, and 21.92% teaching across multiple educational levels. The descriptive statistics are shown in

Table 1. Descriptive statistics of the sample.

Features	n	%
Gender
Male	89	21.92
Female	317	78.08
Age
20–29	90	22.17
30–39	130	32.02
40–49	133	32.76
≥50	53	13.05
Years of teaching experience
<5	117	28.82
5–10	95	23.40
11–20	102	25.12
>20	92	22.66
Educational level
Elementary	116	28.57
Middle school	87	21.43
High school	114	28.08
Multiple levels	89	21.92

.

6.2. Research Instrument and Data Analysis

This study employed a previously validated survey instrument to detect the drivers that affect K–12 IUChatbot, and the analysis was guided by the UTAUT2 model. The survey instrument encompassed the six factors proposed in the research model (Figure 1): PEX, EEX, SIN, FCOs, HMO, and IUChatbot. The UTAUT2 model was extended with the AIL factor adopted from Wang et al. [14]. The research instrument comprised two sections: the first section gathered demographic information about the teacher participants, and the second section consisted of 33 statements distributed across the aforementioned factors that were designed according to a five-point Likert scale ranging from ‘strongly agree’ (5) to ‘strongly disagree’ (1).

Data analysis was conducted using the partial least squares structural equation modelling (PLS-SEM) technique, implemented through the SmartPLS 4.0 software. As outlined by Hair et al. [67], the PLS-SEM method involves a two-step process: First, the measurement model is assessed to ensure reliability and validity. The structural model is then assessed to test and confirm the research hypotheses related to the proposed constructs.

7. Findings

7.1. Measurement Model

At this stage, the model is measured and evaluated based on specific criteria, primarily relying on confirmatory factor analysis (CFA). During this process, validity and reliability are simultaneously verified. Convergent validity is assessed to determine the extent to which the items of a scale align with their corresponding latent variable, which is measured through factor loadings (FLs), which should not be less than 0.7 at a statistical significance level of 0.05 [67]. Additionally, it is important to ensure that the average variance extracted (AVE) for each dimension does not fall below 0.5. Regarding discriminant validity, its role here is to ensure that the factors are distinct from each other, with no overlap or similarity between them. This is confirmed by squaring the AVE value and ensuring that the smallest AVE value is greater than the smallest correlation coefficient in the correlation matrix. For the reliability measurement, both Cronbach’s alpha (CA) and composite reliability (CR) are used, with both values required to be at least 0.7. To assess the model’s fit, we use the normed fit index (NFI), which indicates the model’s contribution and should not be less than 0.90. Additionally, the standardised root-mean-square residual (SRMR) is employed to evaluate the model’s accuracy and its representation of the data, with a value not exceeding 0.080 [67].

Based on the aforementioned steps in the statistical analysis process, it is crucial first to ensure that the items appropriately represent their respective scales, i.e., the FL of each item should exceed 0.70. Accordingly, convergent validity was confirmed (see

Table 2. Analysis of validity and reliability.

Construct	Indicator	FL	CA	CR	AVE	R2
PEX	PEX 1	0.91	0.85	0.86	0.77	-
	PEX 2	0.85
	PEX 3	0.87
EEX	EEX 1	0.85	0.87	0.88	0.79	-
	EEX 2	0.87
	EEX 3	0.81
SIN	SIN 1	0.89	0.83	0.83	0.75	-
	SIN 2	0.89
	SIN 3	0.89
FCOs	FCOs 1	0.88	0.80	0.81	0.71	-
	FCOs 2	0.89
	FCOs 3	0.81
HMO	HMO1	0.90	0.86	0.87	0.78	-
	HMO 2	0.87
	HMO3	0.88
AIL	AIL 1	0.71	0.92	0.93	0.62	-
	AIL 2	0.73
	AIL 3	0.76
	AIL 4	0.81
	AIL 5	0.80
	AIL 6	0.81
	AIL 7	0.82
	AIL 8	0.84
	AIL 9	0.84
IUChatbot	IUChatbot 1	0.92	0.91	0.91	0.84	0.64
	IUChatbot 2	0.93
	IUChatbot 3	0.91

). As shown in Table 2, the scales demonstrated high reliability, with CA and composite reliability values ranging from 0.85 to 0.93. Additionally, the AVE values exceeded 0.5, ranging from 0.62 to 0.84. These findings validate the model’s effectiveness and confirm its reliability in addressing the study’s goals, consistent with the framework proposed by Hair et al. [67].

As illustrated in

Table 3. The correlation matrix, discriminant validity, and model strength.

Constructs	1	2	3	4	5	6	7
1. PEX	0.87
2. EEX	068 **	0.89
3. SIN	0.47 **	0.47 **	0.87
4. FCOs	0.58 **	0.65 **	0.50 **	0.84
5. HMO	0.71 **	0.67 **	0.47 **	0.65 **	0.88
6. AIL	0.65 **	0.74 **	0.51 **	0.64 **	0.71 **	0.79
7. IUChatbot	0.63 **	0.64 **	0.44 **	0.60 **	0.74 **	0.76 **	0.92
NFI value = 0.93 AVE = 0.074 ** Significant correlation at the 0.01 level

, the correlation coefficients ranged from 0.47 to 0.76 and were statistically significant at the 0.01 level. The diagonal elements of the matrix, reflecting the square root of the AVE, exceeded the corresponding correlation coefficients, thereby confirming discriminant validity. The lowest square root value observed was 0.79, surpassing the highest correlation value of 0.76 within the matrix. In terms of model robustness, the NFI was reported at 0.92, and the SRMR stood at 0.072. Both values underscore the model’s precision and fit to the data.

7.2. Structural Model

AIL is further described as encompassing a range of skills that enable effective and critical interaction with AI applications (Figure 2). This stage relies on three key values: the R² value, which quantifies the extent to which the independent variables account for the variance observed in the dependent variable; the t-value; and the standardised regression coefficient (β) [67]. The study model showed that the R² value for teachers’ IUChatbot was 0.64, indicating that 64% of the variance in their intention to use chatbots was explained by the six factors (shown in Table 2).

Table 4. Results for each hypothesis.

H	Independent Variable	Path	Dependent Variables	(β)	t-Values	p-Values
H1	PEX	->	IUChatbot	0.16	1.11	0.27
H2	EEX	->		0.05	0.83	0.41
H3	SIN	->		0.14	** 2.85	0.00
H4	FCOs	->		0.10	0.32	0.75
H5	HMO	->		0.30	* 2.02	0.04
H6	AIL	->		0.43	** 7.03	0.00
* p-value < 0.05, ** p-value < 0.001

and Figure 2 show that AIL (β = 0.43, p < 0.001) has the strongest and most significant positive impact on teachers’ IUChatbot, followed by HMO (β = 0.30, p < 0.05). SIN also shows a significant positive effect (β = 0.14, p < 0.001). However, other factors, such as PEX (β = 0.16, p = 0.27), EEX (β = 0.05, p = 0.41), and FCOs (β = 0.10, p = 0.75), do not have a statistically significant impact on teachers’ IUChatbot. These results highlight the crucial roles of AIL and HMO in shaping teachers’ IUChatbot in educational settings.

8. Discussion and Implications

8.1. Impact of PEX on IUChatbot

The primary hypothesis explored the link between PEX and teachers’ IUChatbot, positing that a positive perception of performance enhancement would lead to higher adoption rates, as proven by many studies [34,52,68]. Contrary to this expectation, the results revealed a non-significant relationship (β = 0.16, t = 1.11, p = 0.27), leading to the rejection of the hypothesis. This suggests that teachers do not necessarily perceive the performance benefits of chatbots as a compelling reason to adopt them. The non-significant impact of PEX might stem from several factors, including the novelty of chatbots in educational contexts and the potential disconnect between perceived utility and actual experience.

Teachers may acknowledge the theoretical benefits of chatbots, such as improved student engagement and streamlined administrative tasks, as highlighted by Venkatesh et al. [13], but they may lack the experiential knowledge to fully appreciate these advantages. These findings suggest a need for increased exposure and training to help teachers understand and experience the tangible benefits of chatbots in enhancing their teaching efficacy. Additionally, the lack of significant influence might be attributed to the variability in teachers’ experiences and comfort levels with technology, which could moderate the impact of perceived performance benefits. As such, teachers may need increased awareness of the benefits of chatbots to form a positive perception of this technology, its adoption, and its effective use [69]. Educational decision-makers play a crucial role in raising awareness of the benefits of interactive chatbots, the reasons for adopting them in education, and the advantages they can provide to optimise and reinforce the educational process. This encouragement can enhance teachers’ awareness of their value [70]. In response to these challenges, educational institutions need to implement robust training programmes that showcase the advantages of chatbots and provide hands-on experience that allows teachers to witness these benefits first-hand. By facilitating workshops, demonstration sessions and pilot programmes, institutions can bridge the gap between perceived and actual utility, thereby enhancing teachers’ readiness to adopt chatbot technologies.

8.2. Impact of EEX on IUChatbot

The second hypothesis explored the effect of EFX on teachers’ IUChatbot, supposing that lower perceived effort would correlate with higher adoption rates. The results, however, did not support this hypothesis, as the relationship was found to be non-significant (β = 0.05, t = 0.83, p = 0.41). This finding indicates that the ease or difficulty of using chatbots is not a significant determinant of teachers’ intentions to integrate these tools into their practice. This result aligns with the findings of [23,34,71]. Several factors might explain this outcome. First, the novelty of chatbot technology in the educational sector could mean that teachers are yet to form strong opinions about the effort required to use such tools effectively. Teachers might use them out of enthusiasm for a new technology, but they have not yet realised the amount of effort and time required for training and effectively integrating them into education to achieve the desired outcomes. Therefore, their intentions were not influenced by this factor. Furthermore, this result might reflect a general resilience or adaptability among teachers who are accustomed to integrating new technologies, suggesting that the perceived effort may not be a critical barrier unless the technology is demonstrably difficult to use.

To mitigate any potential concerns related to effort, educational institutions and technology developers should focus on enhancing the user-friendliness of chatbot platforms. By incorporating intuitive design features, clear instructional materials, and ongoing technical support, developers can reduce the cognitive load associated with adopting new technologies. Institutions and educators should ensure that chatbots are designed to be straightforward and accessible for users, integrate them into current educational environments and practices such as learning platforms, and intensify practical training. In addition, they should offer the necessary support for teachers in managing classrooms, designing educational curricula, creating activities aligned with learners’ interests, and developing development plans, as well as automating the assessment and feedback mechanisms and other educational practices [37,69]. Additionally, providing teachers with access to peer support networks and forums where they can share tips and strategies for efficient chatbot use could further alleviate concerns about effort.

8.3. Impact of SIN on IUChatbot

The third hypothesis addressed the role of SIN in shaping teachers’ IUChatbot, positing that teachers are more likely to adopt chatbots if they perceive that their peers and influential figures within their professional environment support the technology. This hypothesis was supported by the data (β = 0.14, t = 2.85, p = 0.00), confirming that SIN significantly impacts teachers’ decision to use chatbots. This result confirms the findings of several studies [53,59,60,68].

This finding underscores the importance of social dynamics in the adoption of educational technologies. Teachers, like many professionals, are influenced by the behaviours and opinions of their colleagues, supervisors, and institutional leaders. If these key figures endorse and utilise chatbots, it creates normative pressure for others to follow suit [13,68,72]. Educational institutions should harness the power of SIN by creating communities of practice in which teachers can collaborate, share experiences, and learn from one another about integrating chatbots into their teaching. Leaders and early adopters should be encouraged to serve as role models, demonstrate the effective use of chatbots, and advocate for their benefits. Additionally, incorporating chatbot technology into institutional policies and strategic plans can further legitimise its use, making adoption a collective—rather than individual—endeavour.

8.4. Impact of FCOs on IUChatbot

The fourth hypothesis examined whether FCOs, such as infrastructure availability and technical support, are a determinant of teachers’ IUChatbot. The lack of a significant relationship (β = 0.10, t = 0.32, p = 0.75) resulted in the hypothesis being rejected. Several studies in the literature have found the same result [27,57,71,73].

This outcome suggests that FCOs are not a primary concern for teachers when deciding whether to use chatbots. This might be due to teachers’ recognition of the Ministry of Education’s efforts to establish high-tech infrastructure and improve communication networks in most schools to keep up with technological transformation [74]. This could be due to the perception that chatbots do not require extensive resources beyond what is already available. Moreover, teachers may feel confident in their ability to navigate technological challenges independently or with minimal support, reducing the perceived importance of external FCOs.

While FCOs did not emerge as a significant factor, institutions should not overlook the importance of maintaining and upgrading the technological infrastructure that supports the use of educational tools, such as chatbots. Continuous assessment of infrastructure needs and the provision of reliable technical support will ensure that potential barriers to adoption remain minimal. Furthermore, providing customised support and resources that accommodate the unique needs of each teacher can create a more favourable environment for chatbots’ adoption.

8.5. Impact of HMO on IUChatbot

The fifth hypothesis focused on the impact of HMO on teachers’ IUChatbot, proposing that teachers who find using chatbots enjoyable and satisfying are more inclined to adopt them. The findings confirmed this hypothesis (β = 0.30, t = 2.02, p = 0.04), indicating that motivation is a critical driver of the process of technology adoption [27,57]. Teachers who derive satisfaction from using chatbots are likely to perceive them as valuable tools that enhance their teaching experience, thereby increasing their intention to integrate them into their practice [51,54].

To capitalise on the motivational factors driving chatbots’ adoption, developers should focus on creating engaging, user-centric designs that make interaction with chatbots both enjoyable and rewarding. Educational institutions could further enhance motivation by incorporating elements of gamification and providing incentives for teachers who actively use and innovate with chatbots in their classrooms. Additionally, highlighting success stories and case studies of teachers who have effectively used chatbots could inspire others to explore the technology with enthusiasm.

8.6. Impact of AIL on IUChatbot

The sixth hypothesis proposed that AIL would positively influence teachers’ intentions to use chatbots, suggesting that teachers who are knowledgeable about AI and its applications are more likely to adopt chatbot technologies. The data strongly supported this hypothesis (β = 0.43, t = 7.03, p = 0.00), indicating that AIL was the most significant driver of teachers’ IUChatbot in this study.

This finding underscores the critical role that AIL plays in the successful adoption of AI-driven educational tools. Teachers who possess a solid understanding of AI concepts and applications are better equipped to appreciate the potential of chatbots, leading to a higher likelihood of their integration into their teaching practice. The more skills teachers have in using AI tools while considering ethical issues, and the greater their capacity to critically assess and interpret the information generated by AI, the greater their intention to use and accept this technology [14,75]. This result is consistent with the findings of previous research by Ng et al. [31], Zhao et al. [75], and Chai et al. [66], which emphasise the importance of AIL in navigating and utilising AI technologies effectively.

Considering the crucial importance of AIL in the adoption of chatbots, educational institutions should place a strong emphasis on incorporating AIL into their professional development programmes. These initiatives should encompass a broad spectrum of topics, such as the ethical implications of AI, critical assessment of AI tools, and practical applications of AI within the classroom setting. By providing teachers with essential skills and knowledge, institutions can cultivate a culture of innovation and bolster confidence in the use of AI technologies. Furthermore, collaboration with AI experts and continuous updates on AI advancements can keep teachers informed and engaged with the latest developments in educational technology.

9. Conclusions, Limitations, and Future Perspectives

This study’s results highlight the essential factors that influence teachers’ acceptance of chatbot technologies in educational contexts. SIN, HMO, and AIL were identified as significant drivers of teachers’ IUChatbot, whereas PEX, EEX, and FCOs were not found to significantly influence their decisions. These outcomes highlight the nuanced nature of technology adoption and point to the importance of employing a multifaceted strategy to promote the acceptance and integration of new tools.

This study has certain limitations that could influence the interpretation and broader applicability of its findings. One primary limitation is that this study employed a self-reported survey methodology, which inherently carries the risk of response bias. Participants may have provided socially desirable answers or responses that reflect their perceived expectations rather than their actual practices or beliefs. Furthermore, concentrating simply on the UTAUT2 factors and AIL may have resulted in the exclusion of other potentially influential factors that impact the adoption of chatbot technology. Additionally, it is important to consider that the early stages of chatbot technology’s adoption present a unique context that may not fully represent the long-term perceptions and usage patterns of teachers. At this preliminary stage, the novelty of chatbots may lead to heightened curiosity or scepticism, potentially skewing teachers’ responses and intentions either positively or negatively. Additionally, the lack of widespread, in-depth experience with chatbots in educational environments during this period means that teachers’ perceptions are likely based on limited interactions and understanding. This limitation suggests that the findings of this study might not capture how these perceptions and intentions could evolve as the technology matures or becomes more integrated into educational systems, and as teachers gain more familiarity and practical experience with its applications.

Future research could address these limitations by incorporating qualitative methods, such as interviews or case studies, to provide deeper insights into the nuanced experiences of teachers and the contextual factors that influence their adoption of chatbot technologies. Expanding the theoretical framework to include other relevant factors, such as perceived technological anxiety, institutional incentives, and the role of professional development in chatbots’ adoption, could also offer a more comprehensive understanding of the drivers of and barriers to the use of chatbots in education. In addition, future research could consider investigating the impact of demographic factors (age, gender, experience) on the factors affecting teachers’ acceptance of chatbots in education, as well as expanding on the study of AIL aspects (awareness, usage, evaluation, AI ethics) and their impact on teachers’ acceptance of chatbot use. Lastly, future research would benefit from longitudinal studies examining how teachers’ perceptions and intentions change over time as chatbot technology becomes more established and its capabilities become more widely understood.