The Impact of AI Awareness on Employees’ Job Burnout: A Chain Mediation of Perceived Organizational Support and Organizational Commitment

Abstract

Job burnout among university teachers has become an increasing concern in the context of artificial intelligence integration in higher education. As core contributors to teaching quality and academic innovation, university teachers may perceive artificial intelligence as a potential threat to their professional roles, which can heighten burnout and affect educational outcomes. This study investigates the impact of artificial intelligence awareness on job burnout among university teachers and examines the chain mediating roles of perceived organizational support and organizational commitment. A total of 434 university teachers completed validated scales for artificial intelligence awareness, job burnout, perceived organizational support, and organizational commitment. Data were analyzed using SPSS Statistics 26 and the PROCESS macro version 4.0 (Model 6). The results showed that artificial intelligence awareness positively predicted job burnout, while perceived organizational support and organizational commitment were negatively associated with it. Both perceived organizational support and organizational commitment mediated the relationship between artificial intelligence awareness and burnout, and a significant chain mediation effect was observed. These findings indicate that artificial intelligence awareness may increase job burnout among university teachers, but strong perceived organizational support can enhance organizational commitment, jointly reducing this effect. The study clarifies the mechanism linking artificial intelligence awareness to job burnout and provides guidance for mitigating AI-related burnout in higher education.

Keywords

AI awareness perceived organizational support organizational commitment job burnout chain mediation

Introduction

With the advent of the AI era, the significant advantages of AI in cost reduction, efficiency enhancement, innovation promotion, and growth facilitation are profoundly reshaping organizational structures and the career development paths of employees across various industries (Sadeghi et al., 2024; Wu et al., 2024; Wu & Zhang, 2024; Yi et al., 2024). Recent studies indicate that employees’ awareness of AI (AI awareness, AIA) consists of two distinct dimensions: threat appraisal and opportunity appraisal (J. Y. Bai et al., 2025). However, research has shown that employees’ most salient and pervasive psychological responses to AI center on threat appraisal. Threat appraisal refers to employees’ perception that AI may jeopardize job security, diminish the relevance of their skills, or disrupt career trajectories—particularly concerns about being replaced by AI. Compared with opportunity appraisal, these negative experiences more directly predict stress reactions and job burnout (JB). In addition, this perception reflects a form of job-related psychological stress rooted in resource threat, which aligns with both the Conservation of Resources Theory (COR; Hobfoll, 2001; Hobfoll et al., 2018) and the Job Demands–Resources (JD-R) Model (Ab, 2007). According to COR, perceived threats to valued resources (e.g., stable employment or professional identity) are potent triggers of emotional exhaustion. Similarly, the JD-R model identifies threat appraisal as a job demand that consumes emotional energy and contributes to burnout. Given that resource threats generally exert stronger and more immediate psychological effects than potential gains, this study focuses on the threat appraisal dimension of AIA to investigate its impact on job burnout and to examine the mediating mechanisms within a resource-based theoretical framework.

Theories and Hypotheses

Theoretical Basis

The COR conceptualizes stress as arising from the perceived depletion or tangible reduction of valued resources (Hobfoll, 2001; Hobfoll et al., 2018). In this study, AIA is defined as a perception of technological threat—specifically, the risk of job replacement or erosion of professional relevance. This framing aligns with the JD-R Model (Ab, 2007), which categorizes job characteristics into demands and resources. Here, AIA functions as a job demand, representing a psychological stressor that consumes energy and elicits emotional strain. Empirically, Zheng and Zhang (2025) showed that AIA positively predicts job insecurity, which subsequently leads to emotional exhaustion via work–family interference. Similarly, B. J. Kim and Kim (2024) found that AI-induced job insecurity reduces psychological safety and increases knowledge-hiding behaviors.

The JD-R model underscores that job resources, such as perceived organizational support (POS) and organizational commitment (OC), serve as key factors in alleviating the strain caused by job demands. In line with this, POS serves as a contextual resource that restores emotional and psychological reserves. Hou and Fan (2024) found that POS mitigates the detrimental effects of AI-induced stress on employees’ work engagement, confirming its role as a protective factor. Moreover, POS can enhance OC, an attitudinal resource that reinforces employee resilience. Thus, combining COR and JD-R perspectives, this study positions AIA as a perceived threat, while POS and OC represent sequential job resources that help sustain well-being and reduce burnout. This integrated model offers a strong theoretical foundation and directly addresses the reviewer’s concern by justifying AIA as a valid form of job-related stressor.

AIA and JB

JB is a psychological syndrome primarily characterized by stress responses resulting from prolonged work pressure. It typically stems from a mismatch or incompatibility between employee expectations and job demands and is regarded as a persistent, chronic stress state (Maslach, 2003; P. Zeng & Hu, 2024). Maslach and Jackson (1981) classified JB into three dimensions: emotional exhaustion (feeling drained due to the excessive use of emotional resources), depersonalization (manifesting as indifferent or even inhumane reactions toward colleagues and clients), and reduced personal accomplishment (negative self-evaluation of work abilities and achievements). This physically and mentally exhausting state of burnout has serious adverse effects on both organizational development and employee well-being, including increased turnover rates, deteriorating health, decreased work efficiency, reduced organizational citizenship behaviors, and a significant decline in overall well-being (Xie et al., 2024). In recent years, research on the factors influencing JB has primarily focused on both internal individual factors and external environmental factors (P. Zeng & Hu, 2024). Internal factors include personal cognition (such as AIA and OC) and individual characteristics (including personality, empathy, and emotional intelligence; Kong et al., 2021; Parmar et al., 2022; Yue et al., 2022). External environmental factors encompass OC (X. Zeng et al., 2020), work stress (Zhao et al., 2021), and social support (Moisoglou et al., 2024). In the context of the AI era, AIA is considered one of the key triggers of JB. For example, Kong et al. (2021) found in their study on the impact of AIA on hotel employees that AIA not only directly increased employee JB but also exacerbated burnout through a mediating mechanism of reduced OC. Therefore, effectively alleviating the JB caused by AIA has become a major focus of both academic research and practical intervention.

AIA, POS, JB

Among the external environmental factors that alleviate JB, POS has garnered increasing attention as a crucial resource. POS reflects how employees perceive the support provided by their organization, including its care for their needs and recognition of their contributions (Eisenberger et al., 1986). According to the COR, receiving additional resource support can effectively offset the emotional exhaustion caused by resource depletion (Hobfoll, 2001). Therefore, POS can effectively moderate the impact of negative emotions, reduce work-related stress and anxiety, and help alleviate JB levels (Tang et al., 2023; Q. Wang & Wang, 2020; X. Zeng et al., 2020). For instance, Tang et al. (2023) examined the psychological well-being of psychiatric nurses and found a significant negative correlation between POS and JB, meaning that nurses who perceived more support showed fewer signs of burnout. When employees face challenges and stress at work, additional organizational support can serve as a valuable resource that replenishes emotional and social reserves, enabling them to cope more effectively with work demands (Y. Li, 2023; Stinglhamber et al., 2016). When employees perceive AI implementation as a potential threat to their job security, organizational support can serve as a buffer, easing their anxiety and lowering the likelihood of turnover. Nonetheless, current research seldom investigates the mediating role of POS between AIA and JB.

AIA, OC, JB

Similarly, the important role of OC in alleviating JB among employees has also garnered increasing attention. OC reflects how strongly employees identify with their organization and engage in its activities, as well as their willingness to continue working there (Allen & Meyer, 1996). According to Allen and Meyer (1996), OC can be broken down into three key elements: emotional attachment, dependency, and ethical obligation. Affective commitment involves the emotional connection and sense of belonging employees experience with their organization; continuance commitment arises from the perceived personal or financial drawbacks of leaving; normative commitment is based on employees’ feeling of moral duty to stay with the organization, often due to a sense of social duty (N. Li et al., 2022). According to the COR, OC represents an attitudinal resource that can help employees preserve and protect other valuable resources. Higher OC enhances employees’ emotional attachment and loyalty toward the organization, which in turn increases their willingness to invest energy in work and reduces the risk of JB (Chambel & Carvalho, 2022; Drew et al., 2024; X. Wang et al., 2022). For example, X. Wang et al. (2022) highlighted a clear negative link between OC and JB when examining factors affecting employee job performance in China’s district-level healthcare system. Similarly, Chambel and Carvalho (2022) studied the relationship between organizational affective commitment and well-being among contact center employees, finding that affective commitment effectively reduced JB. Furthermore, Kong et al. (2021) confirmed that OC mediates the relationship between AIA and JB. With the introduction of AI, changes in organizational models, and the increase in job instability and uncertainty, employees’ OC has declined. When OC decreases, it commonly leads to lower levels of employee motivation and a rise in JB (Kong et al., 2021).

AIA, POS, OC, JB

According to COR, when employees receive sufficient organizational support, it serves as a valuable resource that helps them maintain and build other resources. Such support enhances employees’ ability to cope with work demands, fosters greater engagement and job satisfaction, and reduces their intention to leave the organization (Caesens & Stinglhamber, 2020; Stinglhamber et al., 2016). OST suggests that POS has a significant positive impact on employees’ OC (Duong et al., 2024; Silva et al., 2022; C. J. Wang, 2022). For example, Duong et al. (2024) investigated the complex relationships among POS, mental health, OC, and nurses’ intention to stay in Vietnam’s healthcare system, finding that POS significantly enhanced nurses’ mental health, strengthened their OC, and increased their intention to remain in the organization. Furthermore, based on the Job Demands-Resources (JD-R) Theory, a lack of job resources or uncertainty around them can cause ongoing energy depletion at work, which decreases OC and may also trigger JB (Bakker, 2015). In contrast, when employees have more resources at work (such as social support from colleagues and supervisors, as well as positive work relationships), they feel more supported and recognized by the organization. As a result, they invest more energy into their work, alleviating concerns about work uncertainty (such as AIA), which not only effectively mitigates JB but also helps enhance employees’ work engagement (Pimenta et al., 2024).

Current Study

Based on the above theories and researches, the impact of AIA on employee JB may be mitigated through the chain mediation of POS and OC. Given the complex effects of AI’s widespread application in the workplace on employees’ psychology and behavior, and in order to fill the research gap in this area, this study systematically explores the mechanism through which AIA influences employee JB. It also provides the first in-depth analysis of the chain mediating roles of POS and OC in this process. Unlike previous studies that mainly examined either direct effects or a single mediator, this model simultaneously incorporates POS and OC to reveal how contextual resources and attitudinal resources interact in sequence to reduce burnout. This chain mediation perspective offers a clearer view of the resource-gain process within the COR framework, highlighting a novel pathway for mitigating AI-related burnout. Building on previous empirical studies, we have developed the theoretical hypothesis model shown in Figure 1 and propose the following hypotheses:

H1: Employees’ AIA is significantly and positively associated with JB;

H2: POS mediates the effect of employees’ AIA on JB;

H3: OC mediates the effect of employees’ AIA on JB;

H4: POS and OC play a chain mediating role in the effect of employees’ AIA on JB.

Figure 1.

Hypothesis model.

Method

Sample and Data Collection

The survey was conducted via the Questionnaire Star online platform (www.Sojump.com) between September and December 2024, and the sample consisted exclusively of faculty members from multiple universities in China. This occupational group was chosen because university teachers perform core tasks such as preparing and delivering lectures, grading assignments, and advising students, which are increasingly supported or partially replaced by AI technologies (e.g., automated grading systems, AI-based course design, intelligent tutoring systems). These developments make them particularly relevant for examining AIA and burnout. A preliminary pilot study was administered to teachers from the intended sample group to ensure that the items were clear, the wording was appropriate, and the overall survey procedure was feasible. Insights gained from the pilot study were applied to optimize the questionnaire before the official data collection began. In the main survey, convenience sampling was adopted, and faculty leaders from various universities distributed the survey information and QR code to faculty work groups via WeChat. All participants were full-time university faculty members, and administrative staff were not included. Each individual received information regarding the research purpose, procedure, and data security prior to submitting online informed consent.

A total of 452 questionnaires were collected for this study. Eighteen invalid questionnaires were discarded, resulting in 434 valid responses (Zhou, 2025). The exclusion of 18 participants’ data was due to two reasons: first, if more than 20% of the questions in a questionnaire were left unanswered, the questionnaire was deemed invalid; second, some participants failed to answer the questionnaire seriously, selecting extreme options (i.e., “Strongly Agree” or “Strongly Disagree”) for more than 80% of the questions. Such response tendencies may introduce bias into the data, potentially reducing the precision of subsequent analyses (D. Wang et al., 2012). Therefore, the final valid sample consisted of 434 responses. The model ensured that no endogenous latent construct was targeted by more than three directional paths. J. F. Hair (2011) noted that, at a 1% significance level, a minimum of 65 participants is necessary for the model to reach an R² of .10. The study utilized 434 samples, significantly exceeding this minimum requirement, which helps to enhance the stability and credibility of the findings. SPSS Statistics 26 was employed for the demographic analysis, with the findings displayed in Table 1. The participants consisted of 128 males (29.5%) and 306 females (70.5%); most participants were between 26 and 45 years old (94.7%); work experience was primarily between 6 and 20 years (89.6%); most held a bachelor’s degree or higher (88.7%); and the salary range was mainly between 5,000 and 15,000 (91.4%).

Table 1.

Demographic Characteristics of the Sample.

Demographic characteristics	Category	Quantity	Percentage
Gender	Male	128	29.5
Gender	Female	306	70.5
Age	18–25	20	4.6
	26–35	349	80.4
	36–45	62	14.3
	46 and above	3	0.7
Work experience	1–5 year	40	9.2
	6–10 year	246	56.7
	11–20 year	143	32.9
	21 year and above	5	1.2
Education level	Associate degree	27	6.2
	Bachelor’s degree	111	25.6
	Master’s degree or above	274	63.1
	Other	22	5.1
Salary range	≤5,000	28	6.5
	5,000–8,000	277	63.8
	8,000–15,000	120	27.6
	≥15,000	9	2.1

Measurement Instruments

The questionnaire is divided into two main sections: the first section collects participants’ demographic information, and the second section gathers self-reported data on various constructs. Validated scales were employed to measure these constructs, with necessary adaptations to fit the specific aims and setting of the research. Beyond collecting basic demographic details, the survey also covered core variables including AIA, POS, OC, and JB. A 7-point Likert scale was employed for scoring, ranging from (1) “strongly disagree” to (7) “strongly agree.”

The AI Awareness Scale, adapted from S. Bai et al. (2024), measures employees’ concerns about being replaced by AI. In adapting the scale, the term “employees” in the original items was replaced with “teachers” to reflect the cultural and occupational context of the education sector, while preserving the original meaning and measurement intent of each item. It consists of four items (e.g., “I believe my job could be replaced by artificial intelligence”). The scale demonstrated good reliability and validity, with a Cronbach’s alpha of .745 and confirmatory factor analysis (CFA) showed satisfactory fit: χ²/DF = 1.871 (≤5), CFI = .995 (≥0.9), TLI = .985 (≥0.9), RMSEA = .045 (≤0.1).

The Perceived Organizational Support Scale, based on the work of Hoa et al. (2020), evaluates employees’ perceptions of the support provided by their organization. It includes 16 items (e.g., “The company assesses my contributions through benefits”). Following Sarstedt et al. (2021), items with standardized outer loadings below 0.70 were examined for removal, and those whose deletion improved composite reliability (CR ≥ 0.70) and average variance extracted (AVE ≥ 0.50) were deleted. Three items were removed based on these criteria, resulting in a 13-item scale, with a Cronbach’s alpha of .920 and CFA showed satisfactory fit: χ²/DF = 1.857 (≤5), CFI = .978 (≥0.9), TLI = .973 (≥0.9), RMSEA = .044 (≤0.1).

The Organizational Commitment Scale, adapted from Alomran et al. (2024), assesses employees’ sense of identification with and loyalty to their organization. It comprises three dimensions: affective commitment, continuance commitment, and normative commitment, with a total of nine items. To reduce measurement redundancy, two items with inter-item correlations exceeding 0.7 were removed. The scale’s Cronbach’s alpha was .865, and CFA showed satisfactory fit: χ²/DF = 4.427 (≤5), CFI = 0.958 (≥0.9), TLI = 0.937 (≥0.9), RMSEA = 0.089 (≤0.1).

The Job Burnout Scale, developed based on Xie et al. (2024), consists of three dimensions: emotional exhaustion, work attitude, and personal accomplishment. It measures the psychological and emotional depletion resulting from prolonged emotional investment and high-intensity work. In total, the scale includes 15 items, following the guidelines of Sarstedt et al. (2021), items with standardized factor loadings below 0.60 were removed, and items with loadings between 0.40 and 0.60 were considered for removal if doing so improved the composite reliability (CR ≥ 0.70) and average variance extracted (AVE ≥ 0.50) while maintaining theoretical consistency. Based on these criteria, six items were deleted, resulting in a final nine-item scale. The Cronbach’s alpha of .877 and CFA showed satisfactory fit: χ²/DF = 4.225 (≤5), CFI = .940 (≥0.9), TLI = .920 (≥0.9), RMSEA = .086 (≤0.1).

The translation process for the scales was conducted collaboratively by two experts: one was a domain expert fluent in both English and Chinese, and the other was a linguist specializing in Chinese language. Experts performed independent translations of the original English scales into Chinese, ensuring accuracy in both meaning and linguistic expression. Through review and discussion, major discrepancies between the two versions were reconciled, resulting in a unified final version. This final version underwent a pilot test with 20 Chinese employees to identify and resolve potential issues arising during the translation process, leading to the completion of the adapted Chinese versions of the scales.

Data Analysis

We used SPSS Statistics 26 to carry out all data analyses. To begin, we calculated descriptive statistics and performed correlation analyses, along with reliability and validity checks, to understand the basic distribution of the data and the performance of each measurement instrument. Prior to hypothesis testing, we examined potential common method variance and assessed multicollinearity to confirm the suitability of the data for multivariate analysis. We then used multiple regression to analyze the associations among AIA, POS, OC, and JB. To further explore the sequential mediation process, PROCESS macro (Model 6) was employed, allowing us to assess whether AIA influences JB through a stepwise effect via POS and OC. The bootstrapping approach with 5,000 samples was used to estimate indirect effects with confidence intervals, ensuring greater accuracy and the stability of the mediation findings.

Results

Common Method Bias Test

We applied the Harman single-factor test to check for common method bias. The exploratory factor analysis extracted 11 factors with eigenvalues above 1, and the first factor explained 36.624% of the variance, remaining below the critical 40% cutoff (Podsakoff & Organ, 1986). Moreover, the results of the latent variable approach for testing common method bias showed that the chi-square difference between the constrained model and the baseline model was 3.334 (df = 1), which did not reach a significant level (p > .05). This indicates that common method bias is not a concern in this research.

Collinearity Test

A collinearity test was conducted on the data obtained in this study. Typically, multicollinearity issues in the model can be assessed using the variance inflation factor (VIF). According to the empirical rule, all constructs should have VIF values below 3.3 (J. Hair et al., 2021). The VIF values for all variables, presented in Table 2, range from 1.702 to 1.753, confirming that multicollinearity does not pose any problem in the model.

Table 2.

Collinearity Test Results.

Independent variable	Tolerance	VIF
AIA	0.583	1.702
POS	0.571	1.753
OC	0.583	1.716

Reliability Assessment of the Measurement Model

Following the recommendations of J. Hair et al. (2021), the reliability of the scales was assessed through measurement model analysis. Reliability indicators included item external loadings and CR, with loadings above 0.7 considered acceptable (J. Hair et al., 2021). As shown in Table 3, all items’ loadings and CR values met the recommended criteria, indicating satisfactory internal consistency reliability. In addition, the AVE values were all greater than 0.5, meeting the criterion for convergent validity.

Table 3.

Reliability and Validity of the Model.

Constructs	Loadings	Cronbach’s alpha	CR	AVE
AIA	0.745	.745	0.839	0.566
	0.752
	0.764
	0.746
JB	0.701	.877	0.901	0.502
	0.710
	0.709
	0.701
	0.707
	0.709
	0.720
	0.707
	0.714
OC	0.725	.865	0.896	0.552
	0.723
	0.770
	0.773
	0.754
	0.732
	0.721
OS	0.709	.920	0.931	0.509
	0.716
	0.722
	0.702
	0.709
	0.708
	0.717
	0.724
	0.709
	0.715
	0.706
	0.713
	0.721

Descriptive Outcomes and Correlation Findings

Based on the descriptive and correlation findings presented in Table 4, significant correlations were found among AIA, POS, OC, and JB (p < .001, N = 434). Specifically, the mean score of AIA is 4.29 (SD = 1.15), which shows a significant positive correlation with JB and significant negative correlations with OC and POS. The mean score of JB is 4.43 (SD = 0.91), which shows significant negative correlations with both OC and POS. The mean score of OC is 4.47 (SD = 1.05), which shows a significant positive correlation with POS.

Table 4.

Descriptive Statistics and Correlation Coefficients of AIA, POS, OC, and JB.

Variables	M ± SD	AIA	JB	OC	POS
AIA	4.29 ± 1.15	1
JB	4.43 ± 0.91	.577***	1
OC	4.47 ± 1.05	−.571***	−.558***	1
POS	4.54 ± 0.87	−.565***	−.582***	−.611***	1

Note. N = 434.

***

p < .001.

Mediation Effect Analysis

After adjusting for factors such as gender, age, years of work experience, education level, and salary, the regression analysis results are presented in Table 5. First, AIA was significantly and negatively associated with POS (β = −.550, p < .001) and OC (β = −.348, p < .001), while it was significantly and positively associated with JB (β = .247, p < .001. Additionally, POS was significantly and negatively related to JB (β = −.210, p < .001). Second, OC was significantly and negatively associated with JB (β = −.321, p < .001). Finally, POS was significantly and positively related to OC (β = .397, p < .001). None of the control variables showed significant effects in any of the models, indicating that the observed relationships are robust to these demographic and occupational factors.

Table 5.

Regression Coefficients for Direct and Indirect Effects of AIA on JB with Control Variables.

Outcome variable	Predict variable	β	SE	t	Bootstrap 95% CI		R ²	F
Outcome variable	Predict variable	β	SE	t	LLCI	ULCI	R ²	F
POS	AIA	−.550	0.042	−12.968***	−0.466	−0.633	.341	40.685
	Gender	−.169	0.105	−1.614	−0.375	0.037
	Age	.107	0.102	1.047	−0.094	0.308
	Years of work experience	−.044	0.079	−0.549	−0.199	0.112
	Education level	−.056	0.079	−0.708	−0.211	0.099
	Salary	−.006	0.098	−0.058	−0.198	0.187
OC	AIA	−.348	0.061	−5.718***	−0.228	−0.468	.425	69.996
	POS	.397	0.058	6.827***	0.282	0.511
	Gender	−.104	0.099	−1.055	−0.299	0.090
	Age	.103	0.082	1.261	−0.058	0.263
	Years of work experience	−.077	0.078	−0.983	−0.230	0.077
	Education level	−.038	0.077	−0.492	−0.188	0.113
	Salary	−.085	0.093	−0.919	−0.268	0.097
JB	AIA	.247	0.054	4.576***	0.141	0.353	.484	73.121
	POS	−.210	0.060	−3.501***	−0.092	−0.327
	OC	−.321	0.053	−6.055***	−0.217	−0.426
	Gender	−.171	0.092	−1.858	−0.351	0.010
	Age	.127	0.088	1.440	−0.046	0.301
	Years of work experience	−.128	0.065	−1.955	−0.256	0.001
	Education level	.053	0.063	0.847	−0.070	0.176
	Salary	−.042	0.074	−0.571	−0.187	0.103

***

p < .001.

Table 6 presents the results of the mediation analysis. First, the total effect of AIA on JB is 0.544, with a 95% confidence interval of [0.459, 0.630]. The direct effect is 0.247, with a 95% confidence interval of [0.141, 0.353], indicating that AIA affects JB both directly and indirectly through mediating variables. Standardized indirect effects were interpreted according to Cohen (2013) benchmarks (small ≈ 0.01, medium ≈ 0.09, large ≈ 0.25) to evaluate their practical significance.

Table 6.

Mediation Effects Analysis of Variables in the Mediation Model.

Effect type	Pathway	Effect value (β)	SE	Percentage (%)	Bootstrap 95% CI		Result
Effect type	Pathway	Effect value (β)	SE	Percentage (%)	LLCI	ULCI	Result
Total effect	AIA → JB	0.544	0.043	100	0.459	0.630	Supported
Direct effect	AIA → JB	0.247	0.054	45.40	0.141	0.353	Supported
Indirect effect	AIA → POS → JB	0.115	0.033	21.14	0.054	0.183	Supported
	AIA → OC → JB	0.112	0.025	20.59	0.069	0.167	Supported
	AIA → POS → OC → JB	0.070	0.017	12.87	0.042	0.108	Supported

Second, when POS is used as a mediating variable, the indirect effect is 0.115, with a 95% confidence interval of [0.054, 0.183], and the mediation effect accounts for 21.14% of the total effect. This value represents a medium-to-large mediation effect, indicating that POS plays a meaningful and practically important role in the relationship between AIA and JB.

Third, when OC is used as a mediating variable, the indirect effect is 0.112, with a 95% confidence interval of [0.069, 0.167], and the mediation effect accounts for 20.59% of the total effect. This represents a medium effect, suggesting that OC is also an important mediator with practical relevance.

Finally, regarding the chain mediating effect of POS and OC, the indirect effect is 0.070, with a 95% confidence interval of [0.042, 0.108], and the chain mediation effect accounts for 12.87% of the total effect. This corresponds to a small-to-medium effect, indicating that although weaker than the single mediators, the sequential mechanism of POS and OC still contributes meaningfully to the overall mediation process.

Discussion

This study investigates the chain mediating role of POS and OC in the relationship between AIA and employee JB, verifying four hypotheses. Specifically, enhancing POS and OC can effectively alleviate the JB caused by high AIA, providing new theoretical support for addressing employee burnout induced by the introduction of AI.

The finding that AIA was significantly and positively associated with employees’ JB (β = .247, medium effect size) is consistent with the findings of S. Bai et al. (2024), who also identified AI-related job insecurity as a driver of burnout. Our result extends prior work by framing this relationship within COR theory, emphasizing the psychological stress employees experience in response to AI-related uncertainty and its impact on burnout. While S. Bai et al. (2024) primarily emphasized workload and task changes, our findings suggest that the psychological threat posed by AI, particularly the uncertainty over future employment, can be an equally important or even more salient predictor of burnout. This difference underscores the need for organizations to address not only operational changes but also the emotional and cognitive resource loss associated with AI adoption. In practice, this means implementing interventions that preserve employees’ sense of stability and control, thereby reducing the risk of burnout in AI-integrated work environments.

POS mediates the impact of AIA on JB, validating H2. The indirect pathway yielded an effect of 0.115 (21.14% of the total effect). This proportion suggests the mediating influence was of a relatively substantial level. Based on the JD-R model, AIA—particularly its threat appraisal dimension—can be conceptualized as a job demand that increases emotional strain, while POS serves as a key job resource that helps buffer this strain. A strong sense of organizational support enables employees to manage the uncertainty brought by AI more effectively, which in turn helps lower emotional exhaustion and burnout. For example, Harunavamwe and Ward (2022) demonstrated that POS buffers the negative impact of technostress on workplace well-being by supporting individuals through emotionally taxing circumstances. This evidence reinforces the protective role of POS in reducing AI-related burnout risk.

OC mediates the relationship between AIA and JB, with the indirect pathway yielded an effect of 0.112 (20.59% of the total effect), this represents a medium to large-scale effect. And consistent with the findings of Kong et al. (2021), validating H3. This underscores the practical relevance of OC in AI-integrated workplaces. This underscores the practical relevance of OC in AI-integrated workplaces. In line with the findings of Chambel and Carvalho (2022), higher OC reflects stronger emotional and psychological investment in the organization, fostering trust and belonging that help counteract the job insecurity triggered by AI adoption. Viewed through the lens of the COR theory, OC represents a valuable resource gain that replenishes depleted emotional and cognitive reserves, supporting employees in maintaining resilience against burnout. Our results extend previous work by indicating that OC also nurtures role identity and a sense of accomplishment (Drew et al., 2024), which together encourage a more constructive perspective toward AI’s role in the workplace and help mitigate its adverse psychological impact.

The results indicate that POS and OC jointly form a sequential pathway through which AIA influences JB, with the chain indirect effect estimated at 0.070, representing 12.87% of the total effect and a small-to-medium effect size. Although the magnitude is smaller than that of the single mediators, the sequential mechanism offers important insight into how organizational resources operate in combination. POS can be seen as an initial resource that fosters OC, which in turn strengthens employees’ willingness and capacity to invest further in their roles, thereby protecting them from burnout. This finding complements earlier studies that examined POS and OC separately, showing that their joint enhancement can produce a cumulative buffering effect against AI-related job insecurity. From a managerial perspective, the evidence suggests that organizations facing AI-driven change should not treat POS and OC as isolated initiatives but should integrate support programs, skill development, and strategies to deepen OC, achieving a synergistic impact on employees’ well-being.

Impact

Theoretical Impact

This study extends burnout research by introducing AIA as a novel psychological job demand within the JD-R model. While traditional burnout literature has emphasized workload, role ambiguity, and emotional labor, technological uncertainty—particularly anxiety about AI replacing human roles—has received less attention. Our findings demonstrate that AIA imposes cognitive and emotional demands on employees, contributing to burnout risk.

The study also enriches the JD-R framework by identifying POS and OC as essential job resources. POS offers external emotional and instrumental support, while OC reflects internal motivation and alignment with organizational values. Both play complementary roles in reducing emotional exhaustion.

Moreover, the chain mediation model reveals that job resources can interact synergistically: POS enhances OC, and together they jointly buffer the negative effects of AIA. This dynamic view highlights how integrated resource pathways can promote employee well-being in AI-integrated workplaces, offering a nuanced extension of the JD-R model.

Practical Impact

This study provides a systematic approach for university administrators to address employee JB in the context of AI technology implementation in higher education. First, the findings indicate that AIA was significantly and positively associated with teachers’ JB. Universities should monitor faculty members’ psychological states and adaptability during the AI integration process in teaching and research. Implementing structured AI transition programs for educators, including phased adoption of AI teaching tools, departmental briefings on potential impacts on teaching loads and academic evaluation, and individualized adaptation plans can help reduce uncertainty and anxiety regarding career development, thereby lowering JB incidence.

Second, the research identifies POS as a crucial intermediary in the link between AIA and JB, offering significant guidance for academic management strategies. Higher education institutions should establish transparent communication channels, such as regular AI pedagogy briefings, faculty consultation forums, and online Q&A platforms, to clarify AI’s scope and its impact on academic roles. They should also organize AI pedagogy workshops to align teaching skills with emerging tools, and adopt participatory curriculum design processes where faculty contribute to AI tool selection and integration into course delivery. Additionally, universities can create peer mentoring programs and discipline-specific AI teaching support groups to enhance faculty members’ sense of control, reducing job insecurity and emotional strain.

Furthermore, this research confirms the important role of OC in mitigating JB. To strengthen OC in higher education, institutions can implement recognition schemes for innovative AI teaching practices, develop academic career advancement pathways incorporating AI-related competencies, and organize interdisciplinary teaching and research collaborations to foster a sense of belonging and professional engagement (S. Y. Kim & Cho, 2022).

Finally, the chain mediation model proposed in this study suggests a multi-level, synergistic intervention strategy. Enhancing POS not only directly improves teachers’ emotional well-being but also fosters AI teaching readiness and psychological safety, strengthening academic community identity. Combining these measures with ongoing faculty well-being assessments, continuous feedback channels, and accessible counseling services can safeguard faculty well-being and teaching performance during AI transitions, achieving both successful AI adoption and sustainable institutional development.

Limitations and Future Research

Although this study reveals the relationships between AIA, POS, OC, and JB, and proposes a chain mediation model, there are several limitations. First, the gender distribution in the sample exhibits a certain imbalance. Among the participants, 128 were male employees, and 306 were female employees, with women accounting for 70.5% of the total sample. This disproportion may be related to the gender distribution characteristics of the industry or job types associated with the study, as female employees might be more concentrated in roles or functions relevant to this research. The gender imbalance may limit the external validity of the results. Future studies should ensure a more balanced gender representation in their samples to improve the generalizability of the findings. Second, as the study is based on a cross-sectional survey, it does not allow for definitive conclusions about cause-and-effect associations in the variables. Future research could utilize longitudinal or experimental approaches to better assess the directionality and causality of the observed effects. Third, the data in this study were mainly collected through self-reports, which may result in biases, including social desirability and memory-recall bias. In particular, the measurement of AIA relied on participants’ self-perceived exposure to AI without independent verification of the actual extent or nature of AI implementation in their institutions. This may have led to discrepancies between perceived and actual AI integration. Future studies could incorporate objective indicators, such as institutional AI adoption records, system usage logs, or third-party evaluations, to validate self-reported data. Fourth, this study did not distinguish participants by academic specialization. Potential differences in AI exposure and attitudes across disciplines (e.g., technical fields involving more automatable tasks and humanistic fields relying more on creativity and emotional interaction) could not be examined, which limits the interpretation and generalizability of the findings. Future research should consider disciplinary differences to enhance the generalizability of the findings. Finally, as the sample in this study was drawn mainly from a single cultural context and consisted of university teachers, the generalizability of the findings is limited. The results are more applicable to work environments characterized by a higher degree of AI adoption and use. Caution is needed when applying these conclusions to other organizational settings or to groups with lower AI involvement. Future research should include samples from different cultural and industry contexts to further test the model’s applicability and enhance the generalizability of the findings.

Conclusion

This study explored how AIA influences employees’ JB and analyzed the chain mediating roles of POS and OC. Based on 434 valid responses, the results confirmed that AIA significantly and positively predicts JB, and that both POS and OC serve as mediators, with a significant chain mediation effect. This study offers new insights into how employees respond to technological uncertainty by identifying organizational support and commitment as key psychological resources that reduce burnout. Practically, organizations can reduce burnout by strengthening POS through transparent communication, targeted training, and inclusive participation, and by enhancing OC through recognition programs, career development, and supportive leadership. However, this study still has limitations, including the gender imbalance in the sample, the focus on higher education employees, and the use of a cross-sectional design that limits causal interpretation. Future research could broaden the sample scope, ensure greater gender and job diversity, and adopt longitudinal or experimental approaches to verify and extend these findings.

Footnotes

ORCID iD

Xingyu Zhou

Ethical Considerations

The researchers confirms that all research was performed in accordance with relevant guidelines/regulations applicable when human participants are involved (e.g., Declaration of Helsinki or similar). This study was approved by the Ethics Committee of Shenyang Normal University (Approval Number: 2024-003).

Consent to Participate

The participants received oral and written information and provided written informed consent before participating in the study.

Author Contributions

Conceptualization: Dongsheng Wang; Methodology: Dongsheng Wang; Formal analysis and investigation: Xingyu Zhou; Writing - original draft preparation: Dongsheng Wang; Writing - review and editing: Dongsheng Wang; Supervision: Xingyu Zhou. All the authors have read and agreed to the published version of the manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The raw data supporting this research are available in the Zenodo repository at https://doi.org/10.5281/zenodo.14915849 (Zhou, 2025).

References

(2007). The job demands-resources model: State of the art. Journal of Managerial Psychology, 22, 309–328. https://doi.org/10.1108/02683940710733115

Allen

N. J.

Meyer

J. P.

(1996). Affective, continuance, and normative commitment to the organization: An examination of construct validity. Journal of Vocational Behavior, 49(3), 252–276. https://doi.org/10.1006/jvbe.1996.0043

Alomran

A. M.

Ahmed

T. S. A.

Kassem

A. M.

(2024). Impact of organizational trust on organizational commitment: The moderating effect of national identity. Cogent Social Sciences, 10(1), 26. https://doi.org/10.1080/23311886.2024.2309712

Bai

J. Y.

Tc Huan

T. C.

Leong

A. M. W.

Luo

J. M.

Fan

D. X. F.

(2025). Examining the influence of AI event strength on employee performance outcomes: Roles of AI rumination, AI-supported autonomy, and felt obligation for constructive change. International Journal of Hospitality Management, 126, 126. https://doi.org/10.1016/j.ijhm.2025.104111

Bai

Zhang

Yao

(2024). Assist me or replace me? Uncovering the influence of AI awareness on employees’ counterproductive work behaviors. Frontiers in Public Health, 12, 1449561. https://doi.org/10.3389/fpubh.2024.1449561

Bakker

A. B.

(2015). A job demands–resources approach to public service motivation. Public Administration Review, 75(5), 723–732. https://doi.org/10.1111/puar.12388

Caesens

Stinglhamber

(2020). Toward a more nuanced view on organizational support theory. Frontiers in Psychology, 11, 476. https://doi.org/10.3389/fpsyg.2020.00476

Chambel

M. J.

Carvalho

V. S.

(2022). Commitment and wellbeing: The relationship dilemma in a two-wave study. Frontiers in Psychology, 13. https://doi.org/10.3389/fpsyg.2022.816240

Cohen

(2013). Statistical power analysis for the behavioral sciences. Routledge.

10.

Drew

J. M.

Sargeant

Martin

(2024). Why do police consider leaving the profession?: The interplay between job demand stress, burnout, psychological distress, and commitment. Policing: A Journal of Policy and Practice, 18, 13. https://doi.org/10.1093/police/paae036

11.

Duong

C. H.

Y. H.

Hoang

T. K. T.

(2024). Perceived organizational support and its impact on nurses’ intention to stay in Vietnam: Psychological well-being and organizational commitment as mediators. Nursing Practice Today, 11(3), 249–258. https://doi.org/10.18502/npt.v11i3.16172

12.

Eisenberger

Huntington

Hutchison

Sowa

(1986). Perceived organizational support. Journal of Applied Psychology, 71(3), 500–507. https://doi.org/10.1037/0021-9010.71.3.500

13.

Hair

Jr Hair

J. F.

Jr Hult

G. T. M.

Ringle

C. M.

Sarstedt

(2021). A primer on partial least squares structural equation modeling (PLS-SEM). Sage Publications.

14.

Hair

J. F.

(2011). Multivariate data analysis: An overview. In Lovric

(Ed.), International Encyclopedia of Statistical Science (pp. 904–907). Springer.

15.

Harunavamwe

Ward

(2022). The influence of technostress, work-family conflict, and perceived organisational support on workplace flourishing amidst COVID-19. Frontiers in Psychology, 13, 17. https://doi.org/10.3389/fpsyg.2022.921211

16.

Hoa

N. D.

Ngan

P. T. H.

Quang

N. M.

Thanh

V. B.

Quyen

H. V. T.

(2020). An empirical study of perceived organizational support and affective commitment in the logistics industry. Journal of Asian Finance Economics and Business, 7(8), 589–598. https://doi.org/10.13106/jafeb.2020.vol7.no8.589

17.

Hobfoll

S. E.

(2001). The influence of culture, community, and the nested-self in the stress process: Advancing conservation of resources theory. Applied Psychology, 50(3), 337–421. https://doi.org/10.1111/1464-0597.00062

18.

Hobfoll

S. E.

Halbesleben

Neveu

J.-P.

Westman

(2018). Conservation of resources in the organizational context: The reality of resources and their consequences. Annual Review of Organizational Psychology and Organizational Behavior, 5(1), 103–128. https://doi.org/10.1146/annurev-orgpsych-032117-104640

19.

Hou

Fan

(2024). Working with AI: The effect of job stress on hotel employees’ work engagement. Behavioral Sciences, 14(11), 20. https://doi.org/10.3390/bs14111076

20.

Kim

B. J.

Kim

M. J.

(2024). How artificial intelligence-induced job insecurity shapes knowledge dynamics: The mitigating role of artificial intelligence self-efficacy. Journal of Innovation & Knowledge, 9(4), 16. https://doi.org/10.1016/j.jik.2024.100590

21.

Kim

S. Y.

Cho

M. K.

(2022). The effect of nurse support programs on job satisfaction and organizational behaviors among hospital nurses: A meta-analysis. International Journal of Environmental Research and Public Health, 19(24). https://doi.org/10.3390/ijerph192417061

22.

Kong

Yuan

Baruch

Jiang

Wang

(2021). Influences of artificial intelligence (AI) awareness on career competency and job burnout. International Journal of Contemporary Hospitality Management, 33(2), 717–734. https://doi.org/10.1108/ijchm-07-2020-0789

23.

Zhang

(2022). Moderated role of social support in the relationship between Job Strain, Burnout, and organizational commitment among operating room nurses: A cross-sectional study. International Journal of Environmental Research and Public Health, 19(17), 11. https://doi.org/10.3390/ijerph191710813

24.

(2023). Relationship between perceived threat of artificial intelligence and turnover intention in luxury hotels. Heliyon, 9(8), e18520. https://doi.org/10.1016/j.heliyon.2023.e18520

25.

Maslach

(2003). Job burnout: New directions in research and intervention. Current Directions in Psychological Science, 12(5), 189–192. https://doi.org/10.1111/1467-8721.01258

26.

Maslach

Jackson

S. E.

(1981). The measurement of experienced burnout. Journal of Occupational Behaviour, 2(2), 99–113. https://doi.org/10.1002/job.4030020205

27.

Moisoglou

Katsiroumpa

Malliarou

Papathanasiou

I. V.

Gallos

Galanis

(2024). Social support and resilience are protective factors against COVID-19 pandemic burnout and job burnout among nurses in the post-COVID-19 era. Healthcare, 12(7), 13. https://doi.org/10.3390/healthcare12070710

28.

Parmar

Channar

Z. A.

Ahmed

R. R.

Streimikiene

Pahi

M. H.

Streimikis

(2022). Assessing the organizational commitment, subjective vitality and burnout effects on turnover intention in private universities. Oeconomia Copernicana, 13(1), 251–286. https://doi.org/10.24136/oc.2022.008

29.

Pimenta

Duarte

A. P.

Simões

(2024). How socially responsible human resource management fosters work engagement: The role of perceived organizational support and affective organizational commitment. Social Responsibility Journal, 20(2), 326–343. https://doi.org/10.1108/srj-10-2022-0442

30.

Podsakoff

P. M.

Organ

D. W.

(1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12(4), 531–544. https://doi.org/10.1177/014920638601200408

31.

Sadeghi

R. K.

Ojha

Kaur

Mahto

Dhir

(2024). Explainable artificial intelligence and agile decision-making in supply chain cyber resilience. Decision Support Systems, 180, 114194. https://doi.org/10.1016/j.dss.2024.114194

32.

Sarstedt

Ringle

C. M.

Hair

J. F.

(2021). Partial least squares structural equation modeling. In Homburg

Klarmann

Vomberg

(Eds.), Handbook of market research (pp. 587–632). Springer.

33.

Silva

Dias

Pereira

da Costa

R. L.

Gonçalves

(2022). Exploring the direct and indirect influence of perceived organizational support on affective organizational commitment. Social Sciences, 11(9), 13. https://doi.org/10.3390/socsci11090406

34.

Stinglhamber

Caesens

Clark

Eisenberger

(2016). Perceived organizational support (Publication Number 978-1-78471-174-0; 978-1-78471-173-3) ://WOS:000411750700025.

35.

Tang

Wang

Zhou

Wang

Zhang

(2023). The relationship between psychiatric nurses’ perceived organizational support and job burnout: Mediating role of psychological capital. Frontiers in Psychology, 14, 1099687. https://doi.org/10.3389/fpsyg.2023.1099687

36.

Wang

C. J.

(2022). Exploring the mechanisms linking transformational leadership, perceived organizational support, creativity, and performance in Hospitality: The mediating role of Affective Organizational Commitment. Behavioral Sciences, 12(10), 17. https://doi.org/10.3390/bs12100406

37.

Wang

Xing

X. H.

X. B.

(2012). The Healthy Lifestyle Scale for University Students: Development and psychometric testing. Australian Journal of Primary Health, 18(4), 339–345. https://doi.org/10.1071/py11107

38.

Wang

(2020). Reducing turnover intention: Perceived organizational support for frontline employees. Frontiers of Business Research in China, 14(1), 16. https://doi.org/10.1186/s11782-020-00074-6

39.

Wang

Chen

Zheng

Zhang

Huang

Birch

(2022). Relationships between job satisfaction, organizational commitment, burnout and job performance of healthcare professionals in a district-level health care system of Shenzhen, China. Frontiers in Psychology, 13, 10. https://doi.org/10.3389/fpsyg.2022.992258

40.

T. J.

Liang

Wang

(2024). The buffering role of workplace mindfulness: How Job Insecurity of Human-artificial intelligence collaboration impacts employees’ work–life-related outcomes. Journal of Business and Psychology, 39(6), 1395–1411. https://doi.org/10.1007/s10869-024-09963-6

41.

T. J.

Zhang

R. X.

(2024). Exploring the impacts of intention towards human-robot collaboration on frontline hotel employees' positive behavior: An integrative model. International Journal of Hospitality Management, 123, 103912. https://doi.org/10.1016/j.ijhm.2024.103912

42.

Xie

Huang

(2024). Intercultural communication competence and job burnout in MNC employees: The mediation role of job stress. Frontiers in Psychology, 15, 10. https://doi.org/10.3389/fpsyg.2024.1339604

43.

R. C.

Gantz

H. Y.

Feldman

S. R.

(2024). Utilizing artificial intelligence technology with emotional intelligence in clinical office visits. Journal of Dermatological Treatment, 35(1), 2374500. https://doi.org/10.1080/09546634.2024.2374500

44.

Yue

Qin

Wang

Nicholas

Maitland

Liu

(2022). Empathy and burnout in medical staff: mediating role of job satisfaction and job commitment. BMC Public Health, 22(1), 13. https://doi.org/10.1186/s12889-022-13405-4

45.

Zeng

(2024). A study of the psychological mechanisms of job burnout: Implications of person–job fit and person–organization fit. Frontiers in Psychology, 15, 8. https://doi.org/10.3389/fpsyg.2024.1351032

46.

Zeng

Zhang

Chen

Liu

(2020). The influence of perceived organizational support on Police Job Burnout: A moderated mediation model. Frontiers in Psychology, 11, 948. https://doi.org/10.3389/fpsyg.2020.00948

47.

Zhao

Liao

Jiang

Ding

(2021). The relationship between teacher job stress and burnout: A moderated mediation model. Frontiers in Psychology, 12, 784243. https://doi.org/10.3389/fpsyg.2021.784243

48.

Zheng

Zhang

(2025). Association between AI awareness and emotional exhaustion: The serial mediation of job insecurity and work interference with family. Behavioral Sciences, 15(4), 12. https://doi.org/10.3390/bs15040401

49.

Zhou

(2025). Raw data_factors influencing career burnout among employees at Shenyang Normal University [Data set]. https://doi.org/10.5281/zenodo.14915849