A Study on Teachers’ Willingness to Use Generative AI Technology and Its Influencing Factors: Based on an Integrated Model

Abstract

The development of new artificial intelligence-generated content (AIGC) technology creates new opportunities for the digital transformation of education. Teachers’ willingness to adopt AIGC technology for collaborative teaching is key to its successful implementation. This study employs the TAM and TPB to construct a model analyzing teachers’ acceptance of AIGC technology, focusing on the influencing factors and differences across various educational stages. The study finds that teachers’ behavioral intentions to use AIGC technology are primarily influenced by perceived usefulness, perceived ease of use, behavioral attitudes, and perceived behavioral control. Perceived ease of use affects teachers’ willingness both directly and indirectly across different groups. However, perceived behavioral control and behavioral attitudes only directly influence university teachers’ willingness to use AIGC technology, with the impact of behavioral attitudes being stronger than that of perceived behavioral control. The empirical findings of this study promote the rational use of AIGC technology by teachers, providing guidance for encouraging teachers to actively explore the use of information technology in building new forms of digital education.

1. Introduction

With the development of artificial intelligence-generated content (AIGC) technology, digital education is presented with a new window of technological opportunity. Under the impetus of digital transformation in education, the creation of new learning spaces where humans and machines coexist, along with new modes of interaction between them, not only provides “intelligent tools” within educational settings but may also introduce new dimensions in the conceptualization of the educational process. AIGC, which is created based on artificial intelligence algorithms, contributes to advancing education toward human–machine collaboration, knowledge breakthroughs, downward compatibility, and iterative progression toward intelligence [1]. For example, AIGC represented by ChatGPT can not only facilitate interactive Q&A sessions but also enable knowledge retrieval, personalized tutoring, homework assistance, language translation, and interactive learning. In the teaching process, it can achieve systematic guidance and precise feedback [2], thereby offering a diverse range of high-efficiency, high-quality, and large-scale intelligent education experiences. Some experts and scholars have stated that AI technology, with AIGC at the forefront, is a crucial component of future education [3]. Others have argued that AIGC provides new momentum and opportunities for “driving educational digitalization” [4]. Therefore, teachers in the intelligent era should remain sensitive to the application of technology in teaching, leveraging its benefits while mitigating its drawbacks. In implementing digital teaching, they should fully utilize technological resources and the functions of digital environments to promote intelligent education and construct new forms of digital education. This is also a reflection of teachers’ digital literacy in action [5]. Observing the development of the information technology era, from education informatization 1.0 to education informatization 2.0, from “Internet+” education to “Blockchain+”, “Metaverse”, and “Intelligent Education”, the development of educational informatization has closely followed the technological opportunities emerging in the trajectory of technological advancement while maintaining sufficient sensitivity to the digitalization process of society. It is evident that seizing the window of opportunity provided by digital technology and understanding the future development and technological foresight of educational digital transformation is not only a demand of the digital transformation era but also a crucial pathway for the digital development of education and teaching. How to seize the opportunity window brought by AIGC technology in the process of educational digital transformation and how teachers can implement digital education and teaching in this context to promote intelligent education is a topic worth consideration and exploration.

In view of this, this study, from the perspective of teachers’ willingness and behavior to accept AIGC technology, constructs a model of teachers’ acceptance of AIGC technology by drawing on the TAM–TPB framework to explore the influencing factors of their acceptance. This is aimed at helping AIGC technology play a more effective and practical role in education, thereby promoting its optimized implementation in the educational process. Understanding the driving factors behind teachers’ use of AIGC technology and the differences across different educational stages remain important issues in the current focus on educational digital transformation. This paper reviews previous studies on the factors influencing teachers’ use of technology, and on this basis, proposes research hypotheses and an integrated model. It adopts a comparative research perspective to explore the differences in the demand and usage of AIGC technology among teachers at different educational stages. Specifically, this study is designed to answer the following two research questions:

Is there a demand difference among teachers at different educational stages for the use of AIGC technology? How do these demand differences influence teachers’ use of AIGC technology?

Which of the three educational stages is more suitable for teachers to use AIGC technology, and why?

By answering these research questions, this study aims to make the following contributions: First, it helps teachers effectively control the extent of technology use during the digital transformation process, clarifying the impact of stage differences and the appropriateness of actual needs. Second, from a practical perspective, the study results will help teachers formulate teaching strategies empowered by AIGC technology, considering both cost and value factors, to better match the learning needs and teaching efficiency optimization of students at corresponding educational stages.

2. Rationale and Research Hypothesis

2.1. Literature Review

• TAM-based study of teachers’ willingness to use technology

The technology acceptance model (TAM) was first proposed by Davis based on the theory of reasoned action (TRA) [6], which mainly explains that the individual’s motivation to use a new technology is affected by its perceived usefulness and perceived ease of use, and predicts to a certain extent the user’s willingness to use it and their behavioral intention toward the technology. Based on the technology use intention and behavioral intention [6], Davis pointed out that the perceived usefulness and perceived ease of use are the key factors in users’ acceptance of information technology. In the field of education, the TAM is widely used in the study of teachers’ acceptance of IT and teaching models [7,8]. However, the model is mainly explored from the perspectives of technical attributes and sense of use, ignoring the influence of social objective factors [9], so there are many studies to expand on the TAM and propose new models; for example, some scholars have explored teachers’ acceptance of new technology with a combined model of the TAM and the TTF (task-technology fit, TTF), and formed a new model [10]. Some foreign researchers have formed the unified theory of acceptance and use of technology (UTAUT) model based on the integration of the TAM and the TRA (theory of reasoned action, TRA) and other theories [11], which provides new perspectives for the study of teachers’ acceptance of information technology. It provides a new perspective for the study of teachers’ IT acceptance [12]. In summary, it can be seen that TAM can be regarded as a core model for exploring the willingness to use technology [13].

2.

A Study of Teachers’ Willingness to Use Technology Based on TPB

The theory of planned behavior (TPB) is inherited from theory of reasoned action (TRA) proposed by Fishbein [14,15] and other scholars, and then expanded and optimized by Ajzen [16], who has retained the “behavioral attitude” and “subjective paradigm” factors and added the factor of “perceived behavioral control” to form a more mature TPB. According to the TPB, behavioral intention is influenced by subjective norms, behavioral attitudes, and perceptual behavioral control, which determine the individual’s behavioral intentions and behaviors [17]. The advantage of the TPB is that it pays more attention to the individual’s objective evaluation of the use of technology and the influence of objective factors, which helps scholars to construct the decision-making process of the individual’s expected behaviors and makes up for the deficiencies of the technology acceptance theory. Currently, the TPB also plays an important role in the study of teachers’ behavioral intentions [18].

2.2. Research Hypotheses

The intention to use AIGC technology among teachers is the outcome variable in the theoretical model of this study. Behavioral intention refers to an individual’s willingness to perform a specific behavior. In this context, teachers’ intention to use AIGC technology represents their willingness to accept it and the extent to which they use AIGC technology. When considering the factors influencing teachers’ intention to use AIGC technology, this study retains the core elements of the TAM (technology acceptance model) and the TPB (theory of planned behavior), while also incorporating perceived cost as an influencing factor into the TAM model, thereby making the model structure more flexible. Accordingly, this study constructs a conceptual model of the factors influencing teachers’ use of AIGC technology based on the integrated TAM/TPB theory. It tests the effects of perceived usefulness, perceived ease of use, perceived cost, subjective norms, behavioral attitudes, and perceived behavioral control on teachers’ use of AIGC technology. The analysis was conducted using a structural equation model, which allowed for a better exploration of teachers’ intention to use AIGC technology and the underlying influencing mechanisms.

• Perceived usefulness, perceived ease of use and perceived cost

Perceived usefulness (PU) reflects the degree to which an individual believes that the use of a specific system improves work efficiency [6]. Davis believes that the user’s behavior in using the system is influenced by behavioral willingness, and behavioral willingness is constrained by individual attitudes and perceived usefulness; the individual’s perceived usefulness to a certain extent, directly affects the user’s behavioral attitudes [6,19]. In teachers’ behavioral willingness to use AIGC technology, perceived usefulness refers to the degree of benefit perceived by teachers who use AIGC technology in their work process. When teachers perceive “valuable benefits” as a result of using AIGC technology, their willingness to use it will increase. Perceived ease of use (PEU) reflects the degree of difficulty or effort an individual perceives in using a specific system to achieve a goal. In the context of teachers’ behavioral willingness to use AIGC technology, perceived ease of use refers to teachers’ self-perceived ease of use in terms of their mastery of AIGC technology and their assessment of the ease of use in terms of the complexity of using the AIGC technology platform as well as their pedagogical transformation. Some scholars have pointed out that ease of use is one of the key variables affecting teachers’ willingness to use technology [20] and found that perceived usefulness and perceived ease of use are determinants of teachers’ willingness to use technology in a study of teachers’ acceptance and adoption of new technology [21]. It can be seen that both perceived usefulness and perceived ease of use have a significant positive effect on teachers’ willingness to accept technology. In addition to this, Davis argues that individuals’ behavioral attitudes are influenced by both perceived usefulness and perceived ease of use [6]. Based on these views, the following hypotheses are proposed:

H1: 

Perceived usefulness significantly influences teachers’ behavioral intention to use AIGC technology;

H2: 

Perceived usefulness significantly influences teachers’ behavioral attitudes toward AIGC technology use;

H3: 

Perceived ease of use significantly influences teachers’ behavioral intention to use AIGC technology;

H4: 

Perceived ease of use significantly influences teachers’ behavioral attitudes toward AIGC technology use.

To study teachers’ willingness to use AIGC technology, in addition to the influence of factors related to perceived usefulness and perceived ease of use, it is also necessary to consider the effort that teachers put into using AIGC technology, i.e., the perceived cost factor. Perceived cost (PC) reflects a comprehensive evaluation of the benefits perceived to be gained by an individual when using a specific system versus the effort, time, and expense incurred [22]. Some studies have indicated that the cost that users face when using a web product affects their willingness to use it [23]. Since AIGC technology is also a network technology, teachers using AIGC technology will incur the costs of energy, time, expenses, and so on in the process of using network technology, so the variable of perceived cost is introduced into the study. If the input cost perceived by teachers in the use of AIGC technology is large, the behavioral attitudes of teachers in using AIGC technology will be affected, and teachers’ willingness to use AIGC technology will also be reduced. Based on this, the following hypotheses are proposed:

H5: 

Perceived costs significantly negatively influence teachers’ behavioral intention to use AIGC technology.

H6: 

Perceived costs significantly negatively influence behavioral attitudes toward teachers’ AIGC technology use.

2.

Effects of behavioral attitudes, subjective norms, and perceived behavioral control on teachers’ willingness to use AIGC technology

The theory of planned behavior suggests that an individual’s willingness to act is thoughtfully planned by the individual and that the individual’s behavioral attitude (BA), subject norm (SN) and perceived behavior control (PBC) play a major role [24]. Behavioral attitudes reflect an individual’s subjective positive or negative feelings toward the acceptance and use of new technology [25]. When teachers hold positive behavioral attitudes toward the use of AIGC technology, it stimulates behavioral willingness to use this technology. Subjective norms are the degree to which individuals are influenced by significant others or groups when making decisions [24]. In teachers’ use of AIGC technology, subjective norms reflect the social pressures that teachers experience when using AIGC technology from other individuals or groups [24]. For example, when teachers are supported by their leaders and colleagues in using AIGC technology, then teachers tend to behave in a supported manner, which further motivates their willingness to use AIGC technology. One study confirms that subjective norms contribute to technology acceptance and willingness to adopt [26]. Perceived behavioral control reflects teachers’ perceived ability to control the consequences of their technology use behaviors and reflects their awareness of their own ability endowment. The stronger the teachers’ perceived behavioral control in using AIGC technology, the more controllable the external factors are and the stronger the teachers’ willingness to use AIGC technology. In the process of teachers’ use of AIGC technology, perceived behavioral control mainly includes teachers’ perceptions of the degree of control, understanding, and application of AIGC technology. Some studies have shown that perceptual behavioral control has a significant role in promoting teachers’ behavioral willingness to teach information technology [27]. Obviously, when teachers are skilled in the operation and application of AIGC technology and other processes, their behavioral willingness to use AIGC technology will also be stronger. To sum it up, subjective norms, behavioral attitudes, and perceived behavioral control will have a positive effect and high explanatory power on behavioral intentions [27]. Therefore, based on the TPB model, the following hypotheses are proposed:

H7: 

Behavioral attitudes significantly and positively influence teachers’ behavioral intention to use AIGC technology;

H8: 

Subjective norms significantly positively influence teachers’ behavioral intention to use AIGC technology;

H9: 

Perceived behavioral control significantly and positively influences teachers’ behavioral intention to use AIGC technology.

3.

Theoretical integration and modeling

The TAM model and the TPB are both developed on the basis of the theory of rational behavior, and there is a certain degree of complementarity, which improves the explanatory and predictive power of the theory used alone [28]. As early as 1995, some scholars, such as Taylor, integrated the two models to achieve theoretical compatibility and complementarity [29], and were able to explore the factors of individual behavioral intentions more deeply. Many scholars have conducted research using the integration theory and achieved good results [30,31]. With the development of the information technology era, the research of this theory on teachers’ behavioral intention of information technology is also becoming more and more prominent, and it has become an important theoretical basis for the effective analysis of teachers’ behavioral intentions toward information technology teaching [32], the acceptance of blended teaching [33], the willingness of transferring the ability of information technology teaching [34], and the intention of teachers’ use of information technology [35,36], and so forth. Studies have shown that the TAM/TPB integration theory has stronger explanatory power than the respective theories used alone [37].

In addition, according to research in the field of psychology, when individuals perceive the implementation of a specific system to be easy to achieve, they also perceive the consequences of this behavior to be closer to their own expectations [38]. Conversely, it can be inferred from this study that when teachers perceive that the use of AIGC technology is easy to implement, they will perceive that the more controllable and convenient the AIGC technology is, the closer the consequences of the behavior will be to their own expectations, which will increase their behavioral willingness to use AIGC technology. When an individual’s perceived usefulness of implementing a specific system is higher, his or her motivation to implement it and confidence in overcoming difficulties will also increase [39]. Based on this, this study added two paths of “perceived ease of use-perceived behavioral control” and “perceived usefulness-perceived behavioral control” to the model, put forward the following hypotheses, and a model of teachers’ willingness to use AIGC technology based on the TAM–TPB framework was constructed (Figure 1).

H10: 

Perceived usefulness significantly and positively affects teachers’ perceptual behavioral control.

H11: 

Perceived ease of use significantly and positively affects teachers’ perceptual behavioral control.

3. Research Methodology

3.1. Operational Definition and Measurement of Model Variables

This study examines teachers’ perceived ease of use of AIGC technology in terms of their perceptions of how to use AIGC technology, the precautions to be taken in using AIGC technology, and the difficulty of using the tools and platforms of AIGC technology; teachers’ perceived value in terms of the efficiency, quality, and favorable conditions for work and teaching felt during the use of AIGC technology; teachers’ perceived usefulness in terms of the effort, time, and money spent in the use of AIGC technology; teachers’ perceived cost in terms of the effort, time, and money spent in the use of AIGC technology; teachers’ behavioral attitudes toward AIGC technology use in terms of teachers’ evaluations, attitudes, and liking of AIGC technology; support for the use of AIGC technology in their schools, including participation and degree of satisfaction to examine teachers’ subjective norms toward the use of AIGC technology; teachers’ perceived behavioral control toward the use of AIGC technology in terms of teachers’ mastery of the use of AIGC technology, degree of knowledge reserved, and degree of integration of the technology; and teachers’ behavioral willingness to use AIGC technology in terms of the direction of assisted teaching and research, personalized teaching, and automated assessment. For the variables involved in the study, they were all measured in the form of questionnaires. The study designed and developed the scale “Survey of Teachers’ Behavioral Intention to Use AIGC Technology” (

Table 1. Design and sources of observational variables.

Latent Variable	Typology	Content of Variable Measurement	Source of Variables
Perceived usefulness (PU)	Causal variables	PU1: The use of AIGC technology can improve teaching and learning PU2: The use of AIGC technology can improve work efficiency PU3: The use of AIGC technology can help me to work together on my tasks PU4: Using AIGC technology makes it easy for me to access the knowledge I want.	Davis [6], Moon & Kim [40], Chen &Tan, 2004 [41], Huang & Lin [42] Venkatesh & Davis [43]
Perceived ease of use (PEU)	Causal variables	PEU1: I understand the information about AIGC technology and how it is used PEU2: I think it is convenient and easy to use AIGC technology PEU3: I understand the rules of AIGC use and how it is used	Davis [6], Brown, et al. [44], Wu &Wang [45], Huang & Lin [42];
Perceived cost (PC)	Causal variables	PC1: Using AIGC technology would waste too much of my energy PC2: Using AIGC technology will cost me too much money PC3: Using AIGC technology would take too much of my time	Venkatesh & Davis [43];
Subjective norms (SN)	Causal variables	SN1: School teachers agree with my use of AIGC technology SN2: School leaders support my use of AIGC technology SN3: Students are receptive to my use of AIGC technology	Ajzen [46];
Behavioral attitudes (BA)	Causal variables	BA1: I am satisfied with the use of AIGC technology BA2: I support the use and promotion of AIGC technology in my school. BA3: I support regional governments in advancing AI education practices on the ground	Ajzen [46];
Perceptual behavioral control (PB)	Causal variables	PBC1: I was able to balance the synergy between AIGC technology and education PBC2: I have a high level of expertise to support the use of AIGC technology PBC3: I am able to master AIGC technology and utilize it appropriately in my teaching job	Ajzen [46], Hale, et al. [47];
Behavioral willingness (BL)	outcome variable	BL1: I am willing to use AIGC technology for assisted teaching and research BL2: I am willing to incorporate AIGC technology for individualized instruction BL3: I am willing to utilize AIGC technology for automated reviews	Venkatesh [43], Gao Furong [48];

) based on the established hypothetical model, drawing on related studies and the literature analysis. The scale was standardized in the form of a five-point Likert scale, with 1 indicating “completely disagree” and 5 indicating “completely agree”. The numerical value of the latent variables was realized in the form of a scale, based on which the interrelationships among the latent variables were examined for the purpose of empirical analysis.

3.2. Sample Selection and Data Collection

Teachers selected for this study were required to have some knowledge of AIGC technology and the questionnaire was distributed using a combination of stratified and random sampling. The questionnaire survey was conducted in Chifeng City, and the total number of teachers who participated in the survey was 457. Among them, 149, or 32.6%, were from high schools, and 148, or 34.4%, were from universities. Overall, the surveyed teachers in this research were more evenly distributed in terms of gender, age, and school, and the composition of the sample met the diversity and representativeness of the study.

3.3. Data Analysis Tools

This study borrowed SPSS 23.0 and Smart PLS 3.0 software for data analysis and constructing structural equation modeling to explore the effect relationship among factors and to effectively predict teachers’ behavioral intention to use AIGC technology.

4. Results

4.1. Data Quality Analysis

• Measurement model confidence tests

The reliability test detects the internal consistency of the dimensions. The coefficient is between 0 and 1. The higher the value of the test result, the higher the reliability. In basic research, the reliability coefficient should reach 0.80 to be accepted, and in exploratory research, the reliability coefficient should reach 0.70 to be accepted; the reliability coefficients between 0.70 and 0.98 are considered to be high, and it is generally considered that a reliability coefficient of 0.6 or less is not reliable, and it is necessary to redesign the scale or try to re-collect the data and analyze it again [23]. The composite reliability of this study based on the standardized Cronbach’s alpha coefficient reached 0.877. The Cronbach’s alpha coefficients based on sub-grouped data from elementary and middle school, high school, and college were 0.857, 0.861, and 0.906, respectively, which are all above the acceptable standard of 0.7. The reliability of each latent variable is shown in

Table 2. Data Reliability and Validity Tests.

Variable Name	Variable Code	Standardized Load Factor			Cronbach’s α			AVE			CR
		Primary Middle School	Congrats! (on Passing an Exam)	College	Primary and Middle School	Congrats! (on Passing an Exam)	College	Primary Middle School	Congrats! (on Passing an Exam)	College	Primary Middle School	Congrats! (on Passing an Exam)	College
Perceived usefulness	PU1	0.948	0.942	0.95	0.965	0.964	0.947	0.905	0.902	0.907	0.974	0.973	0.975
	PU2	0.95	0.964	0.961
	PU3	0.957	0.947	0.952
	PU4	0.95	0.946	0.946
Perceived ease of use	PEU1	0.959	0.956	0.96	0.959	0.954	0.939	0.925	0.915	0.912	0.974	0.97	0.969
	PEU2	0.963	0.953	0.945
	PEU3	0.963	0.96	0.961
Perceived cost	PC1	0.829	0.862	0.801	0.868	0.887	0.872	0.793	0.817	0.771	0.92	0.93	0.91
	PC2	0.930	0.937	0.925
	PC3	0.909	0.91	0.904
Subjective norm	SN1	0.8	0.805	0.931	0.852	0.844	0.916	0.68	0.722	0.86	0.861	0.885	0.948
	SN2	0.988	0.777	0.931
	SN3	0.65	0.955	0.92
Attitude	BA1	0.902	0.918	0.911	0.894	0.908	0.945	0.825	0.844	0.829	0.934	0.942	0.936
	BA2	0.914	0.905	0.914
	BA3	0.909	0.933	0.907
Perceptual behavioral control	PBC1	0.922	0.966	0.734	0.896	0.888	0.955	0.828	0.751	0.72	0.935	0.897	0.884
	PBC2	0.947	0.971	0.801
	PBC3	0.86	0.613	0.99
Willingness to act	BL1	0.921	0.918	0.934	0.91	0.904	0.908	0.847	0.838	0.844	0.943	0.94	0.942
	BL2	0.926	0.917	0.914
	BL3	0.915	0.912	0.909

, from which it can be seen that the Cronbach’s alpha coefficients of each latent variable are higher than 0.80, showing that the questionnaire has a high reliability.

2.

Model validity tests

After an exploratory factor analysis using SPSS 23.0, the total KMO = 0.962, sig = 0.000, and the KMO values obtained according to the three types of grouping of schools were 0.948, 0.944, and 0.965, and the test coefficients were all greater than 0.5 and the significance was less than 0.001, which made them suitable for factor analysis. Validity tests usually include both convergent validity and discriminant validity. Convergent validity refers to the similarity test for the same traits, and the results should have a high correlation; discriminant validity refers to the difference indicator between different traits, and the results should have a low correlation.

• (1)

  Convergent validity

The convergent validity of the scale was measured using three indicators: the standardized loading coefficients of the question items, the combined reliability (composite reliability, CR) of the latent variables, and the average variance extracted (average variance extracted, AVE); the results are shown in Table 2. As seen in Table 2, the standardized loading coefficients are all above the acceptable range of 0.6, which indicates that the factor loadings of the model are ideal. The combined reliability (CR) coefficients were higher than 0.8, indicating that the internal consistency of the model is good. The average variance extracted (AVE) was also higher than the minimum standard of 0.5, indicating that the model has a better convergent validity.

• (2)

  Distinguishing Validity

Distinctive validity is used to test the relative independence of the latent variables and is measured using the trivariate cross-loading method, the Fornell–Larcker criterion (FLC) and the heterotrait-monotrait ratio (HTMT). The study began with a test using the Fornell–Larcker criterion, in which the correlation coefficients between the latent variables were compared with the square root of the AVE, and the results showed that although the square root of the AVE for each latent variable was greater than the correlation coefficients of the other latent variables, there were cases in which the correlation coefficients of some of the latent variables were larger. To ensure the validity of the data, the HTMT ratio was further examined, and the results showed that, except for perceived usefulness and perceived ease of use, which were 0.873, all the other data were less than the recommended value of 0.85, as shown in

Table 3. Distinctive validity test of HTMT ratio.

	Subjective Norm	Perceived Cost	Perceived Ease of Use	Perceived Usefulness	Perceptual Behavioral Control	Attitude	Willingness to Act
Subjective norm
Perceived cost	0.288
Perceived ease of use	0.263	0.784
Perceived usefulness	0.237	0.750	0.873
Perceptual behavioral control	0.119	0.168	0.200	0.240
Attitude	0.179	0.719	0.806	0.833	0.250
Willingness to act	0.220	0.656	0.811	0.802	0.299	0.778

. And when the concepts of the constructs are similar, the threshold of the HTMT ratio can be relaxed to 0.90 [49], so it can be shown that the differentiation validity of this study is good and suitable for subsequent analysis.

3.

Model Fit Test

“Model fit”, i.e., the degree of matching between the data and the model, is an important influence on the results of the study. In this study, we analyze the standardized root mean square residual fitness (SRMR) of the model based on the existing studies. SRMR is an absolute fit index, which represents the difference between the observed model and the predicted model [50]. By examining the correction index of the overall model, the SRMR value of the model was obtained as 0.042, which is less than the threshold requirement of 0.08. The SRMR values of 0.071, 0.055, and 0.057 were obtained for the three subgroups of elementary and middle school, high school, and university, respectively, which are all less than 0.08, indicating that the model has a good degree of fitness.

4.2. Analysis of Differences in School Segments

Teachers’ behavioral intention to use AIGC technology under the TAM–TPB framework was analyzed using structural equation modeling (SEM) with partial least squares regression (PLS), and the standardized path coefficients of the grouping model are shown in Figure 2, Figure 3 and Figure 4 (*, **, and *** in the diagram indicate p < 0.05, p < 0.01, and p < 0.001, respectively). In the three subgroup models, the latent variables of perceived ease of use, perceived usefulness, behavioral attitudes, and perceived behavioral control have direct or indirect effects on teachers’ willingness to use AIGC technology, while the effect of perceived cost on teachers’ willingness to use AIGC technology is insignificant, as shown in

Table 4. Effects of latent variables on teachers’ behavioral intention to use AIGC technology.

Latent Variable	Intermediary Variable	Direct Effect			Intermediary Effect			Aggregate Effect
		Primary Middle School	Congrats! (on Passing an Exam)	College	Primary and Middle School	Congrats! (on Passing an Exam)	College	Primary Middle School	Congrats! (on Passing an Exam)	College
PEU	PBC, BA	0.246	0.307	0.062	0.018	0.04	0.224	0.264	0.347	0.286
PU	BA	0.466	0.332	0.384	0.008	0.063	0.176	0.474	0.395	0.56
BA		-	-	0.373			-	-	-	0.373
PBC		-	-	0.239			-	-	-	0.239

.

• Perceived Ease of Use Analysis

The data analysis shows that in all three subgroup models, elementary and middle school, high school, and university, teachers’ perceived ease of use significantly and positively affects their behavioral attitudes, i.e., the easier teachers perceive the use of AIGC technology, the more positive their attitudes are toward using AIGC technology. The data analysis reveals that there are differences among the three groups, with perceived ease of use having the strongest effect on teachers’ behavioral attitudes toward using AIGC technology in the university subgroup (0.375), followed by high school (0.310), and the weakest effect in the elementary and middle school subgroup (0.277). In the university subgroup, teachers’ perceived ease of use similarly influenced teachers’ perceived behavioral control (0.354) and behavioral attitudes (0.375), and at the same time, perceived behavioral control and behavioral attitudes significantly influenced teachers’ behavioral willingness to use AIGC technology. In contrast, in the elementary and middle school and high school subgroups, teachers’ behavioral attitudes and perceived behavioral control did not significantly influence teachers’ behavioral intention to use AIGC technology. Therefore, only in the university subgroup, perceived ease of use with behavioral attitudes or perceived behavioral control as a mediator variable had an indirect effect (0.224) on the behavioral intention of university teachers’ AIGC technology use. In contrast, in the elementary and middle school and high school subgroups, the direct effect of perceived ease of use on their behavioral intention to use AIGC technology was mainly reflected (0.246, 0.307).

2.

Perceived usefulness analysis

From the results of the analysis, it can be seen that perceived usefulness has a significant positive effect on both behavioral attitudes and behavioral willingness, and the more teachers perceive that AIGC technology brings convenience and help to their own teaching and learning, the stronger their attitudes and willingness to use AIGC technology will be. Among them, high school teachers’ perceived usefulness had the strongest effect on their behavioral attitudes toward AIGC technology use (0.499), followed in descending order by the university subgroup (0.315) and the elementary and middle school subgroup (0.287). The perceived usefulness of teachers in the elementary and middle school subgroup had the strongest effect on their behavioral intention to use AIGC technology (0.466), with the university subgroup (0.384) and high school subgroup (0.332) acting in descending order. Therefore, perceived usefulness has a direct effect on teachers’ behavioral intention to use AIGC technology on the one hand, and on the other hand, it has an indirect effect on teachers’ behavioral intention to use AIGC technology with behavioral attitude as a mediating variable. And the most significant effect was on the university subgroup (0.176).

3.

Behavioral Attitude Analysis

The analysis of the three-group model showed that the behavioral attitudes of the teachers in the university subgroup had a significant positive effect on their behavioral intentions, i.e., the more positive the behavioral attitudes of the university teachers toward the use of AIGC technology, the stronger their behavioral intentions toward the use of AIGC technology. This also proves the establishment of the mediating effect within the model, i.e., university teachers’ perceived usefulness and perceived ease of use of AIGC technology are mediated by behavioral attitudes on their willingness to use AIGC technology. In the elementary and middle school and high school subgroups, teachers’ behavioral attitudes did not have a significant effect on behavioral willingness. The reason for this is that the current regional promotion of AI education in elementary and middle schools and high schools faces many realistic dilemmas such as a lack of a proper system, resources, collaboration, and teachers’ qualifications [51], which impedes the popularization and development of AI education in elementary and middle school and high school stages in China; in addition to this, the difference in teaching between university teaching and elementary and middle school teaching is also a manifestation of the difference in the grouping model. Compared to the university’s more open knowledge system and complex assessment standards, primary and secondary schools present more fixed textbooks and teaching materials, unified assessment standards, and fixed teachers and classes, so teachers of primary and secondary schools are affected by the habit of forming an inertial dependence on the traditional teaching system and are more inclined to fixed patterns and routines in teaching [52]. Therefore, from this perspective, the demand for AIGC technology among primary and secondary school teachers is lower than that of university teachers, who are academically free.

4.

An Analysis of Subjective Norms and Perceived Behavioral Control

It was found that teachers’ behavioral intention to use AIGC technology was not affected by subjective norms in any of the three subgroups. This is similar to the findings of some researchers [27,53,54]. The reason for this may be related to the teachers’ inherent teaching environment and familiar teaching format, whose use or non-use has less impact on teachers’ subsequent teaching; in addition to this, external factors such as a lack of informatized instructional support and experience may also have a significant relationship with this result. Meanwhile, perceived behavioral control significantly affected teachers’ behavioral intention to use AIGC technology only in the university subgroup (0.239), and teachers’ behavioral intention to use AIGC technology in the elementary and middle school and high school subgroups were not affected by perceived behavioral control. The reason may be that perceived behavioral control mainly includes teachers’ own perceptions of the use of AIGC technology and external perceptions that promote or hinder the use of AIGC technology, and compared with the more academically free universities, teachers in elementary and middle schools and high schools teach according to the existing rules and standards [52], and their own needs for the use of AIGC technology are relatively low. It was also found that limited external resources and teachers’ curiosity about AIGC technology may also contribute to the limited influence of their perceived behavioral control on behavioral intentions.

5.

Perceived cost analysis

It was found that teachers’ behavioral intention to use AIGC technology was not affected by perceived cost in any of the three subgroups. In terms of explanatory factors, it can be seen that PC3 scored the highest among the three variables of perceived cost, i.e., the “time” factor scored 3.03, and the “energy” factor of PC1 and the “money” factor of PC2 had mean values of 2.77 and 2.89, respectively. The mean values of the “energy” factor of PC1 and the “money” factor of PC2 are 2.77 and 2.89. As mentioned earlier, the scale is assigned values from one to five, with three being the median value, so only the “time” factor slightly exceeds the median value. To some extent, this reflects the fact that teachers are not affected by the “energy” and “money” consumed by the use of AIGC technology, while the “time” factor is higher than the other two factors, but not significant in the grouping model. The “time” factor, although higher than the other two factors, did not significantly explain teachers’ perceived costs in the grouping model. It is clear that the perceived cost factor has limited influence on teachers’ willingness to use AIGC technology.

5. Discussion

This paper draws on the TAM–TPB framework in the field of social psychology to construct a model of teachers’ behavioral willingness to use AIGC technology, focusing on analyzing the similarities and differences in the willingness to use the technology in three different subgroups, elementary and middle school, senior high school, and university, which is basically in line with the expectations of the previous paper. The study found that the teachers’ demand for and behavioral attitudes toward the use of AIGC technology varied among the different subgroups of the school year, and was affected by the value of the use of the technology, the cost, and the teacher’s behavioral willingness to use the AIGC technology. This was influenced by multiple factors, as described below:

• In the process of teachers’ use of AIGC technology, their behavioral willingness was mainly influenced by four factors: perceived usefulness, perceived ease of use, behavioral attitude, and perceived behavioral control. Among them, perceived usefulness directly influenced teachers’ willingness to use AIGC in the three subgroups, with the strongest direct effect in primary and junior high schools; at the same time, it also indirectly influenced teachers’ willingness to use AIGC through the intermediary effects of perceived behavioral control and behavioral attitude in some subgroups. It can be seen that the more beneficial AIGC technology is to teaching and educational development and the more teachers perceive the usefulness of AIGC technology, the stronger their behavioral attitudes or perceptual behavioral control and the stronger their willingness to use it.
• Perceived ease of use also affected the willingness to use the technology in different groups of teachers in the form of both direct and indirect effects. The paths of “perceived ease of use–behavioral attitude–behavioral intention” and “perceived ease of use–perceived behavioral control–behavioral intention” indirectly influence the willingness to use the technology. At the university level, the two paths of “perceived ease of use–behavioral attitude–behavioral intention” and “perceived ease of use–perceived behavioral control–behavioral intention” were used to indirectly influence behavioral intention. The more teachers knew about the rules and ways of using AIGC technology, the greater the perceived ease of use and subsequently, the greater the willingness to use AIGC technology. However, across the three subgroups, the total effect of perceived ease of use was relatively weak compared to perceived usefulness.
• Perceived behavioral control and behavioral attitude directly affected teachers’ willingness to use AIGC only in the university subgroup, and the effect of behavioral attitude was higher than that of perceived behavioral control. That is, in the university subgroup, the more positive the teachers’ attitudes toward the use of AIGC technology, the stronger their willingness to use it; in addition, the university teachers’ willingness to use AIGC technology in the teaching and learning process will be stronger when they perceive that the use of AIGC technology in the teaching and learning process is more controllable.
• Teachers in different school segments had slightly different preferences for the use of AIGC technology. Teachers in elementary, junior high, and senior high schools did not have a high demand for the use of AIGC technology, and the influence of the working environment and external factors also limit the use of AIGC technology by teachers in this school segment; in addition to this, teachers in this school segment mostly tried AIGC technology with a curiosity mentality, which leads to the fact that they can still perceive the perceptual behavioral control, subjective norms, and behavioral attitudes presented by AIGC technology in teaching and learning. Teachers in the university subgroup were more likely to use AIGC technology and work with it, and were more likely to use it to support teaching and research.

6. Conclusions

6.1. Theoretical Significance

The proper and appropriate use of AIGC technology by teachers is a crucial prerequisite for ensuring the positive development of AIGC technology in the educational field and the successful digital transformation of education. This study contributes in several specific ways by organizing and analyzing the current state of teachers’ use of AIGC technology and the influencing factors.

First, existing research has widely focused on the impact of AIGC technology on teachers’ work [55], but it has overlooked the mechanism of subjective behavioral intentions, where teachers are the main agents. Only by focusing on teachers’ subjective behavioral intentions can the empowering role of technology be better realized. This study addresses the factors influencing teachers’ individual behavioral intentions, particularly the longitudinal evaluation of teachers’ needs and value judgments, further advancing the theory of teachers’ use of AIGC technology.

Second, the study results indicate that neglecting the differences in educational stages when using AIGC in teaching is unreasonable. The model constructed in this study compares the differences in the use of AIGC technology among teachers in elementary, secondary, and higher education. It reveals differences in needs and cognition among different educational stages, which, in turn, lead to varying behaviors among teachers when using AIGC technology. This finding suggests that teachers’ cognition and the different needs of various educational stages play an essential role in their use of AIGC technology. Therefore, through the TAM–TPB model’s comparison across three educational stages, this study highlights the varying effects of different variables on teachers’ use of AIGC technology, providing a reference for teachers to use AIGC technology appropriately according to their respective educational stages.

Additionally, this study integrates the TAM and TPB into the TAM–TPB model, enriching the comprehensive understanding of teachers’ use of AIGC technology. Specifically, it analyzes both the explicit impact of the value and cost of using technology from a technical perspective and the implicit impact from the psychological needs of teachers. This addresses the gap in previous research that overlooked the implicit influencing factors related to teachers as the main subjects. The study thus provides a theoretical basis for future researchers to develop different technology-empowered educational strategies tailored to different educational stages and teachers’ psychological needs.

6.2. Recommendations

Information technology empowerment contributes to the realization of the goal of improving quality and balancing the development of education [56], in which AIGC technology, as a new track of information technology development, will provide strong support for information technology empowerment in education form upgrading. Teachers’ willingness to use AIGC technology is a key factor. In view of the above analysis, this study puts forward the following suggestions for promoting teachers’ rational use of AIGC technology, motivating teachers to actively explore IT-enabled education development, and promoting digital transformation of education.

First, in the pursuit of fostering teachers’ proactive engagement with AIGC technology, tailored guidance strategies for educators across different academic levels are essential. For university instructors, showcasing exemplary cases of AIGC technology integration in education is paramount, encouraging the ongoing exploration of novel technological tools to enrich learning experiences. Emphasizing the judicious and lawful utilization of AIGC technology, educators are urged to harness its capabilities to foster precise, personalized, and intelligent learning environments, thereby nurturing talents equipped for future societal advancements. Conversely, for primary and secondary school teachers, emphasis lies in accentuating the role and promotion of AIGC technology, coupled with the implementation of digital competency training initiatives. These efforts aim to cultivate teachers’ comprehension and acceptance of the transformative impacts of AIGC technology, fostering a culture of human–machine collaborative intelligence in education. Additionally, attention is directed toward optimizing teachers’ perceived usefulness and ease of use of AIGC technology, facilitating the exploration of innovative teaching methodologies and the seamless integration of AIGC technology to enhance instructional practices. By prioritizing educators’ awareness of the utility and accessibility of AIGC technology, this approach seeks to cultivate a pedagogical landscape characterized by efficacy and adaptability.

Secondly, teachers should grasp the scale and methods of technology-enabled education when using AIGC technology, so as to help intelligent education with smart education. Smart education refers to the use of modern educational technology to facilitate the quality development of education, while intelligent education refers to the use of human–computer synergy to create changes in the teaching process and thus promote the development of learners [57]. The study found that most teachers believe that “it is easy to obtain the required knowledge using AIGC technology”, i.e., under the influence of AIGC technology, the ease of obtaining knowledge contrasts with the difficulty of obtaining wisdom. Therefore, the utilization of the advantages of intelligent education to help smart education, and the cultivation of students’ critical thinking and independent thinking, become the value of shaping human beings in a smart society. From the aspect of technology use, it is necessary to understand the rules and forms of technology introduction to be able to accurately judge the degree of technology introduction in the teaching process, and to avoid the phenomenon of uncontrolled teaching caused by technology abuse or irregular use; from the aspect of one’s own growth, it is necessary to concentrate on the work that cannot be realized by technology, such as the cultivation of students’ emotion and will, ideology and politics, morality and ethics, new types of thinking, etc. From the aspect of teaching, it is necessary for the teachers to accurately position the relationship between human beings and technology, not to use technology as a mere tool for load reduction, and to avoid technology over-reliance leading to burnout. From the aspect of environment construction, the goal must be to build a new modern education system that is ubiquitous, flat, digital, and skillful; to promote the good integration of digital ecology and education and teaching; and to trigger the innovative development of human–computer collaborative education so as to realize the optimal development of intelligent education to help smart education.

Finally, increasing teachers’ willingness to use AIGC technology requires a comprehensive design and improvement in the external environment that frames the improvement of teachers’ digital literacy. The study found that subjective norms and perceived costs have no significant effect on teachers’ willingness to use AIGC technology, and the research hypothesis of this variable is not supported. This suggests that teachers’ willingness to use AIGC technology tends to be more toward rational judgment, and the people or teams around them have limited influence on their use. Therefore, teachers should not be forced to use AIGC technology in the form of “orders”, “documents”, etc., but should be encouraged by publicity and guidance to look at the changes brought by AIGC technology to education in a cautiously optimistic attitude and to do a good job of risk prediction and prevention. AIGC technology has been used in the education sector for many years. Although AIGC technology has a wide range of prospects and potential in the field of education, a number of uncertainties may also bring certain risks to the education system [58]. Therefore, it is necessary to strengthen teachers’ understanding and training, avoid a high degree of dependence on data, and improve their own ability to make choices, make judgments, and think independently, so that teachers can confidently and effectively utilize the technology to assist in education and avoid risks.