Understanding AI Technostress and Employee Career Growth from a Socio-Technical Systems Perspective: A Dual-Path Model

Abstract

The rapid advancement of Artificial Intelligence (AI) has profoundly transformed organizational systems, reshaping how employees interact with technology and adapt to changing work environments. However, the systemic mechanisms through which AI-induced technostress influences employee career growth remain insufficiently understood. Grounded in a socio-technical systems perspective, this study conceptualizes organizations as adaptive systems where technological, organizational, and human subsystems dynamically interact. We propose a dual-path framework that distinguishes between challenge-related technostressors (a resource-gain process) and hindrance-related technostressors (a resource-loss process), elucidating how AI-related pressures can simultaneously foster and hinder career development. Furthermore, employee resilience and organizational AI support are incorporated as systemic moderators that modulate the intensity of these effects within the human–AI–organization system. Using two-stage survey data from 326 matched pairs of employees and supervisors, results largely support the proposed model, with some pathways showing marginal significance. The findings reveal that AI challenge-related technostressors stimulate proactive adaptation and skill development, whereas hindrance-related technostressors generate anxiety and insecurity, thereby impeding growth. This research extends systems theory by demonstrating how technostressors function as an emergent property of human–technology interactions and provides actionable insights for designing more adaptive and resilient socio-technical work systems.

1. Introduction

The rapid advancement of Artificial Intelligence (AI) has been reshaping not only individual work practices but also the broader socio-technical systems within organizations [1,2,3]. From process automation to intelligent decision support, AI technologies are reconfiguring the interdependencies between people, technologies, and organizational structures. These changes extend beyond efficiency and productivity; they redefine feedback loops between human behavior, technical artifacts, and managerial systems, thus, altering employees’ task characteristics, skill requirements, and career development trajectories [4].

However, the integration of AI into organizational systems introduces systemic tensions. While technological innovations generate new opportunities and learning potentials, they simultaneously create new uncertainties, coordination challenges, and psychological burdens—collectively referred to as technostressors [5,6,7]. As a systemic phenomenon, technostress emerges from dynamic interactions between individual cognition, technological complexity, and organizational design. Its effects are, therefore, not purely individual but propagate across the socio-technical system through behavioral and emotional feedback mechanisms.

Technostress has long been recognized as a critical component of the human–technology interface in organizations [2,8]. In the AI era, however, technostress manifests with greater complexity and duality. On one hand, challenge-related technostress can stimulate employees to learn, adapt, and acquire new skills, enhancing their capacity to thrive within evolving systems [9,10]. On the other hand, hindrance-related technostress may produce anxiety, insecurity, and diminished motivation, thereby weakening the adaptive capacity of both individuals and the wider organizational system [11,12]. This duality aligns with the challenge–hindrance stressor framework, which conceptualizes stressors as having both growth-oriented and constraint-oriented effects within complex systems [7,13,14].

Despite growing scholarly interest, several systemic research gaps remain. First, prior studies have primarily examined general information systems or digital tools [15,16], while the stressors emerging from AI adoption—characterized by intelligent autonomy and algorithmic opacity—remain underexplored. Second, although existing research (e.g., Califf et al., 2020) [7] has explored the positive aspects of technological pressure, studies examining its systematic impact on professional growth remain insufficient in the specific context of AI. Third, the dual mechanisms of challenge and hindrance stressors have often been discussed conceptually but seldom empirically validated using systemic modeling approaches such as chain mediation analysis or moderated structural models.

Accordingly, this study investigates AI-induced technostress from a socio-technical systems perspective, addressing three central questions: (1) Does AI-induced technostress exhibit both challenge and hindrance mechanisms within organizational systems? (2) How do these mechanisms propagate through psychological and behavioral subsystems to influence employees’ career growth? (3) How do organizational supports and individual resilience interact as system moderators that amplify or buffer these mechanisms?

The contributions of this study are threefold:

• Theoretical contribution: This study reframes AI technostress as a systemic phenomenon emerging from human–technology–organization interactions. By embedding the challenge–hindrance framework into a socio-technical systems perspective, it extends the theoretical boundary of stress and career development research.
• Empirical contribution: Using multi-source, two-stage survey data and structural equation modeling, this study systematically validates the dual pathways (resource gain and resource loss chains) through which AI-induced technostressors affects career growth, thus, revealing feedback and mediation structures within the system.
• Practical contribution: The findings provide actionable insights for AI-driven organizations. By designing supportive systems that balance technological demands with learning resources and emotional support, organizations can enhance systemic resilience and sustain employee career growth amid digital transformation.

In sum, this study views AI technostress not merely as an individual psychological response but as an emergent property of complex socio-technical systems. It contributes to understanding how human adaptability, technological design, and organizational support interact to shape systemic outcomes in the age of intelligent transformation.

2. Literature Review

2.1. Technostress in the Era of Artificial Intelligence as a Systemic Phenomenon

Technostress has long been recognized as a central construct in information systems and organizational behavior research, defined as the stress individuals experience due to the adoption and use of information technologies [1,2,8]. Earlier studies predominantly viewed technostress as a source of strain—linked to reduced performance, satisfaction, and psychological well-being [5,7,11]. Subsequent research, however, emphasized its dual nature, suggesting that technology-induced stress may also trigger adaptive learning, innovation, and skill enhancement under appropriate conditions [4,9,10].

Although stress fundamentally originates from individual appraisals [17], in the AI era, this individual response is deeply embedded within the feedback loops of organizational systems. AI technologies, characterized by autonomous decision-making and predictive analytics, not only reshape task structures but also redefine socio-technical interactions between humans and intelligent systems [3,7]. Recent system-oriented literature emphasizes that technostress should no longer be viewed in isolation, but as an emergent outcome of feedback loops across human, technological, and structural subsystems [18,19,20].

This systemic orientation draws from foundational systems theory, which posits that organizational behaviors emerge from interdependent components in dynamic interaction [21,22]. Historical cases of socio-technical change—such as Trist and Bamforth’s study of coal mining mechanization—already demonstrated how new technologies can produce unexpected stress due to misalignment between social and technical subsystems [23]. This systems-based view provides a more dynamic understanding of how AI transforms work environments and employee adaptation patterns [24,25].

2.2. The Challenge–Hindrance Stressor Framework as a Dual System Pathway

The challenge–hindrance (C–H) stressor framework provides a valuable lens to understand how stress operates within organizational systems. Challenge stressors act as positive feedback loops that stimulate growth, learning, and achievement, while hindrance stressors function as negative feedback loops that constrain adaptation and lead to system inefficiency [14,26]. Recent studies have extended this framework to technostress, suggesting that technology-induced pressures may operate as either adaptive (challenge) or maladaptive (hindrance) system responses [15,27].

However, the application of this dual-path framework to AI-induced technostress remains limited. AI technologies differ from traditional IT systems by introducing both cognitive enrichment and existential uncertainty—thus, generating competing systemic forces that simultaneously enhance and disrupt employee performance. This study aims to address the gap regarding that it remains unclear whether AI-related stress follows the same dual-system pathways and how these pathways translate into long-term career development outcomes. Recent empirical studies demonstrate that perceived algorithmic control, for example, can trigger both job engagement and withdrawal behavior depending on whether workers appraise the AI system as a challenge or a threat [20,28]. These patterns support the recursive feedback structures emphasized by systems theory. Specifically, supervisor ratings were used to measure the dependent variable, Career Growth, to structurally separate the independent and dependent variables and minimize common method bias in the systemic model. Specifically, employees’ differentiated responses do not mark the end of the interaction; rather, they act as new input signals that feed back into the system, forming a dynamic evolutionary loop of ‘evaluation-behavior-re-evaluation’.

2.3. Technostress and Career Growth Within Socio-Technical Systems

Career growth, broadly defined as employees’ perceived opportunities for skill development, advancement, and self-realization, is deeply embedded in socio-technical systems [29,30]. From a systems perspective, career development reflects the degree of alignment between the individual subsystem (skills, motivation, resilience) and the organizational-technical subsystem (technology, structure, and learning climate). Challenge stressors can activate adaptive loops that enhance learning and proactive behavior, facilitating system-level renewal and innovation [31]. In contrast, hindrance stressors can generate negative loops that deplete psychological resources, induce withdrawal, and reduce the system’s overall adaptive capacity [32,33].

Recent research has shown that AI-induced technostress can significantly undermine psychological safety and increase emotional strain, which in turn suppress career motivation and innovation behavior [20,34]. These effects are compounded by feedback mechanisms, where stress accumulates over time to create systemic inertia or disengagement [18].

Existing studies linking technostress and career-related outcomes—such as employability and innovation—often adopt a linear cause–effect approach [12,13]. However, the dual nature of AI-induced technostress suggests that these relationships are recursive, interdependent, and context-sensitive, consistent with systems theory’s emphasis on feedback and adaptation [24,35].

Therefore, career growth should be understood not as a static outcome but as a dynamic process shaped by continuous human–AI interactions within complex organizational ecosystems [22,25].

2.4. Mediating and Moderating Mechanisms in System Dynamics

To explain the variability in system outcomes under AI-induced stress, researchers have highlighted mediating and moderating mechanisms that regulate systemic equilibrium. Psychological mediators such as self-efficacy, learning orientation, and resilience convert challenge stress into adaptive learning, while contextual moderators such as job resources, organizational AI support, and leadership style buffer the negative effects of hindrance stressors [2,4,6,7].

From a systems-theoretic standpoint, these mediating and moderating factors serve as stabilizing or amplifying subsystems that govern feedback strength within the overall socio-technical structure [21]. When buffering capacity (e.g., resilience, support) is strong, the system maintains equilibrium and converts stress into adaptive energy. When such capacity is weak, feedback becomes destructive, leading to resource loss and systemic degradation [19,34].

Daily-level research further reveals that technostress can trigger a spillover effect—today’s exhaustion leads to next-day behavioral withdrawal—thus, confirming the existence of feedback propagation in technostress processes [18]. Similarly, leadership ethics and perceived AI transparency have been shown to buffer negative effects by restoring psychological safety [34].

However, empirical research validating these dynamic mechanisms in the AI context remains limited. Most studies treat technostress as a unidimensional variable rather than as an emergent system phenomenon with nonlinear feedback effects. Thus, an integrative framework that combines the challenge–hindrance perspective with socio-technical systems theory is urgently needed to explain how AI-induced stress shapes career growth through interconnected psychological and organizational processes [24,25].

3. Hypotheses and Research Model

3.1. Theoretical Foundations: A Systems Perspective

Artificial intelligence (AI) is fundamentally reshaping contemporary organizations, altering work structures, decision-making processes, and the conditions under which employees pursue career development. While AI presents opportunities for skill enhancement, innovation, and adaptive learning, it simultaneously introduces complexity, uncertainty, and stress that may hinder employees’ growth trajectories. Understanding these dual effects requires a systemic lens that integrates individual, technological, and organizational components and captures the feedback loops and interactions among them. To this end, we adopt two complementary theoretical perspectives—the Transactional Theory of Stress and the Job Demands–Resources (JD–R) model—within a socio-technical systems (STSs) framework. Together, these frameworks illuminate how AI-induced technostress functions as both a challenge and a hindrance within the organizational system, affecting employee career outcomes.

3.1.1. Transactional Theory of Stress

The transactional theory of stress conceptualizes stress as a dynamic, system-like interaction between the individual and their environment [36]. Stress emerges not solely from external conditions but from the appraisal processes through which employees evaluate task demands, available resources, and potential outcomes. Within this framework, stressors are categorized as either challenge stressors, which stimulate growth and achievement, or hindrance stressors, which obstruct goal attainment and deplete motivation [26].

Applied to AI, this perspective emphasizes that the systemic impact of technology depends on employees’ perceptions and adaptive responses. Challenge-related AI stressors—such as opportunities to acquire new competencies or solve complex problems—activate positive feedback loops that enhance proactive behaviors and career development. In contrast, hindrance-related AI stressors—such as perceived role ambiguity, algorithmic complexity, or inadequate support—trigger negative feedback loops that amplify insecurity, anxiety, and impediments to career growth.

3.1.2. Job Demands–Resources (JD–R) Model

The JD–R model complements the transactional perspective by emphasizing the structural balance of demands and resources within the socio-technical system [37]. Job demands, including the continuous adaptation to AI systems and cognitive workload, impose physiological and psychological costs on employees. Job resources—such as training, managerial support, and resilience—buffer these demands and can activate motivational feedback loops that enhance engagement, learning, and career growth.

Within the AI context, this dual pathway aligns with systemic principles: when organizational and personal resources match technological demands, AI becomes a growth-enabling element of the socio-technical system. Conversely, misalignment between demands and resources produces reinforcing loops of stress, decreased performance, and hindered career progression.

3.1.3. Integrating the Dual Perspectives

By integrating the transactional theory of stress and the JD–R model within a socio-technical systems perspective, AI-induced technostress is conceptualized as a dynamic, system-level phenomenon. Its effects depend simultaneously on cognitive appraisals, feedback interactions between employees and technology, and the availability of organizational resources. This integration highlights the dual pathways through which AI can either enhance or impede career growth and provides the theoretical justification for modeling challenge- and hindrance-related stressors as distinct but interacting subsystems within the organizational system.

3.1.4. Hypotheses Development

Existing research indicates that the impact of technology-induced stress on employees is multifaceted, encompassing both negative and positive dimensions. On the negative side, technology use can force employees to process excessive information, leaving insufficient time for exploration and innovation, which detrimentally affects individual job performance [38], work quality [39], and productivity [40,41,42]. Additionally, the insecurity stemming from technostress has been found to amplify employee deviant behaviors [43]. Conversely, on the positive side, Saleem et al. [44] identified that information technology can function as a positive stressor, enhancing individual performance. Consequently, given the widespread application of AI, it is logical to infer a significant correlation between AI-induced technostress and individual career growth.

Career growth represents the velocity at which an individual moves toward valued positions, primarily encompassing knowledge capability development, progress in career goals, and opportunities for promotion. Hassan I. Bailout (2009) argues that individuals with high organizational commitment will actively seek challenging work to pursue greater career growth opportunities [45]. Career growth is, thus, the result of the interaction between individual effort and organizational support.

Recent empirical research supports the proposed systemic and dual-perspective framework by examining the opposing psychological mechanisms triggered by AI. Chang et al. [46] explicitly demonstrate that AI-driven technostress acts as a double-edged sword. Supporting the positive pathway, Zhang et al. [20] found that when technostress is appraised as a challenge, it fosters work engagement and drives innovative behavior. Similarly, Ding (2021) proposes that viewing AI as a challenge generates high levels of engagement; employees may proactively upgrade their professional capabilities to mitigate AI-induced insecurity [47,48], thereby facilitating professional development. AI technology can also assist employees in generating creative ideas, an effect particularly significant for skilled workers [49].

Conversely, illustrating the negative pathway, Wang and Zhou [50] provide evidence that heightened AI awareness can function as a hindrance stressor. Recent studies indicate that negative AI awareness inhibits organizational commitment, professional competence, and job satisfaction [51,52], while promoting work overload, job burnout, turnover intention, and depression [53,54]. These adverse effects erode the psychological resources necessary for development, which ultimately obstructs sustainable career growth. These studies collectively substantiate the existence of the proposed positive and negative feedback loops.

Based on this theoretical and empirical foundation, we propose the following hypotheses:

Hypothesis 1 (H1a).

AI challenge-related technostress is positively related to employee career growth.

Hypothesis 1 (H1b).

AI hindrance-related technostress is negatively related to employee career growth.

These hypotheses establish the foundation for modeling the dual pathways of AI technostress in Section 3.2 and Section 3.3, emphasizing systemic interactions among technological, individual, and organizational components.

3.2. AI Challenge-Related Technostress: A Systems Perspective

Drawing on Transactional Theory of Stress, the Job Demands–Resources (JD–R) model, and a socio-technical systems lens, we argue that AI challenge-related technostress operates as a dynamic element within the organizational system, exerting both direct and mediated influences on employees’ career growth. Within this systemic framework, challenge-related AI stressors act as positive stimuli that activate feedback loops of skill acquisition, proactive behavior, and career advancement, while their effects are contingent on both organizational (AI support) and individual (employee resilience) resources.

3.2.1. Mediating Role of AI Personal Utility and Proactive Career Behaviors

AI personal utility reflects how the use of AI satisfies employees’ confidence, independence, and enjoyment at work, representing a positive cognitive and emotional appraisal of technology adoption. Within the JD–R and systems frameworks, AI personal utility functions as a critical job resource. When employees encounter AI challenge-related technostress, features of AI-based information systems are perceived as opportunities to improve individual skills, tasks, and work life, making employees feel more satisfied and competent [55].

This positive appraisal motivates employees to improve their work experience by making active changes to better adapt to and utilize AI technology for workflow optimization and efficiency enhancement [56]. Employees with high AI challenge awareness evaluate this stress as an opportunity for higher achievement; to benefit from career development opportunities, they are more prone to altering task boundaries by undertaking valuable work, thereby generating higher levels of task engagement [57].

Furthermore, high work autonomy—a core component of AI personal utility—facilitates the pursuit of personal career growth goals. Individuals with high independent self-construal define themselves through internal attributes and value personal goals [58] and personal growth [59]. Since career growth is essentially a process of self-concept realization [60], high autonomy empowers employees to commit to their career development goals. This career commitment acts as an intrinsic driver, guiding employees to engage in career-oriented proactive behaviors to approximate their career goals [61].

From a systemic perspective, employees actively engage in human–AI interactions, modifying processes and behaviors to leverage AI for individual and organizational gains. Such proactive engagement constitutes a reinforcing feedback loop: challenge perception → resource utilization → behavior change → career growth.

Hypothesis 2 (H2a).

AI challenge-related technostress positively affects AI personal utility.

Hypothesis 2 (H2b).

AI challenge-related technostress positively influences employees’ career growth through the mediation of “AI personal utility → proactive career behaviors.”

3.2.2. Moderating Role of Organizational AI Support

Organizational AI support constitutes a structural resource within the socio-technical system, shaping employees’ capacity to leverage AI effectively. Theoretically distinguishable from individual traits, it functions as an external job resource provided by the organization. Timely IT support assists employees in learning and using technology, while effective encouragement stimulates interest, thereby mitigating the negative impact of stressors. In a work environment surrounded by peer evaluation, employees assess expected gains and losses before exhibiting innovative behaviors like using new technology [62].

Specifically, organizational support can facilitate the acquisition of AI knowledge—employees’ perception of what they know about AI [63]—which serves as a critical personal resource enabling them to cope with challenging stressors. Adequate support provides technical knowledge, guidance, and a positive climate that reinforces adaptive behaviors and the perception of AI as a developmental tool. Within the system, organizational support strengthens the positive feedback loop between challenge stressors and AI personal utility, enhancing the downstream effects on proactive career behaviors and career growth.

Furthermore, we posit that the mediating pathway is contingent upon this support. Specifically, when organizational AI support is high, employees possess greater willingness and resources to utilize new AI tools; they feel more confident that AI usage will enhance their capabilities and performance, leading to elevated levels of AI personal utility. This heightened utility subsequently triggers proactive career behaviors and facilitates career growth. Conversely, when support is low, employees’ positive attitudes toward AI diminish due to insufficient coping resources, suppressing the translation of stress into utility and subsequent proactive actions.

Hypothesis 3 (H3a).

The positive effect of AI challenge-related technostress on AI personal utility is moderated by organizational AI support, such that the relationship is stronger under high organizational AI support.

Hypothesis 3 (H3b).

The mediating effect of “AI personal utility → proactive career behaviors” between AI challenge-related technostress and career growth is moderated by organizational AI support, such that the mediation is stronger when organizational AI support is high.

3.2.3. Moderating Role of Organizational AI Support: An External Resource

Within the Job Demands–Resources (JD-R) framework, Organizational AI Support functions as a critical exogenous job resource. It represents the structural and contextual aid provided by the organization, such as technical training, IT infrastructure, and managerial encouragement [2]. Unlike individual traits, this form of support operates at the systemic level, modifying the environment in which employees interact with AI. By providing necessary external tools, it reduces the cognitive burden of challenge-related stressors, thereby strengthening the perceived utility of the technology.

3.2.4. Moderating Role of Employee Resilience

Employee resilience primarily refers to the ability of employees to utilize resources to continuously adapt and develop at work, even when facing challenging environments [64]. Research indicates that although resilience is partly genetic, it is not a stable trait; rather, it can be developed in environments that foster adaptability [65]. Unlike organizational support which is exogenous, resilience represents an internal personal resource that enables employees to buffer stress from within.

Employees exhibiting resilient behaviors—being proactive, adaptive, and utilizing networks [64,66,67]—are likely to view their organization and work more positively because they possess the competence to manage challenges and navigate the workplace [68]. Consequently, resilient employees are better equipped to interpret AI-induced demands as opportunities, activating personal resources to perceive higher levels of AI personal utility.

Furthermore, we posit that resilience regulates the indirect pathway to career growth. Specifically, when employee resilience is high, employees are more willing to use AI and can adapt to technical pressures faster and better; this leads to a more significant elevation in AI personal utility, which subsequently generates more proactive career behaviors and fosters career growth. Conversely, when resilience is low, employees possess weaker adaptive capabilities and lower confidence in using AI; consequently, the enhancement of AI personal utility diminishes, leading to fewer proactive behaviors and impeding career growth.

Hypothesis 4 (H4a).

The positive effect of AI challenge-related technostress on AI personal utility is moderated by employee resilience, such that the relationship is stronger when resilience is high.

Hypothesis 4 (H4b).

The mediating effect of “AI personal utility → proactive career behaviors” between AI challenge-related technostress and career growth is moderated by employee resilience, such that the mediation is stronger when resilience is high.

3.3. AI Hindrance-Related Technostress: A Systems Perspective

AI hindrance-related technostress reflects the perception that AI technologies introduce obstacles, uncertainty, and constraints within the organizational system. From a socio-technical systems perspective, these stressors disrupt the balance between job demands and resources, triggering feedback loops that amplify job insecurity and workplace anxiety, ultimately impeding career growth.

Mediating Role of AI-Related Job Insecurity and Workplace Anxiety

Job insecurity refers to employees’ concerns regarding the stability of their employment and future career development, encompassing fears of job loss, deteriorating working conditions, or reduced opportunities [69]. Research indicates that the application of AI increases employees’ concerns about job substitution [70], and technological changes in the workplace significantly affect employees’ adaptability. Within a socio-technical framework, AI implementation modifies work processes, skill requirements, and job characteristics, creating systemic perturbations that may be appraised as threats. Employees perceive these changes as potential losses of valued resources, activating a reinforcing loop: AI hindrance stressors → job insecurity → workplace anxiety → impaired career growth.

Certain individual differences amplify the sensitivity to these systemic perturbations. For instance, individuals with neurotic or obsessive tendencies tend to interpret environmental conditions—such as the requirement for 24/7 connectivity—as threatening stressors. They are prone to anxiety and paranoia, feeling insecure if they miss important information, and are disturbed by blurred boundaries between work and non-work domains. Furthermore, individuals with low self-efficacy may evaluate increased work demands as threats.

Hypothesis 5 (H5a).

AI hindrance-related technostress positively influences AI-related job insecurity.

According to cognitive appraisal theory, job insecurity triggers negative evaluations, leading to workplace anxiety. The uncertainty inherent in job insecurity hinders individuals from effectively using coping strategies [36]. AI-induced insecurity questions the very existence of the industry and the job itself [71], and is negatively correlated with employee attitudes such as organizational commitment and willingness to stay [71,72].

Workplace anxiety represents the emotional and cognitive response to perceived or actual loss of resources within the organizational system. According to the Job Demands–Resources (JD–R) model, anxiety functions as a sustained job demand that continuously consumes individual energy and resources, negatively affecting work- and career-related cognitive decisions. Since career growth largely depends on the enhancement of professional capabilities and the realization of career goals, this resource depletion is critical.

When AI technostress leads employees to perceive their career resources as threatened, it generates strong feelings of insecurity and anxiety regarding potential job loss or the inability to adapt to new requirements. This emotional state triggers individual self-protection mechanisms, causing employees to stagnate and avoid innovative behaviors or risks. Such conservative coping strategies are antithetical to the requirements of organizational innovation and personal career growth, ultimately impairing their work performance and professional development.

Hypothesis 5 (H5b).

AI hindrance-related technostress negatively affects employees’ career growth through the “AI-related job insecurity → workplace anxiety” pathway.

3.4. Integrative Research Model: A Systems Framework

Building on the preceding discussion, we propose a conceptual model that situates AI-induced technostress within a socio-technical systems perspective, emphasizing the dual and interacting pathways through which challenge and hindrance stressors influence career growth. AI challenge-related stressors function as reinforcing loops that enhance skills, proactive behaviors, and career outcomes, whereas AI hindrance-related stressors operate as balancing loops that elevate job insecurity and anxiety, constraining professional development.

The model incorporates mediating mechanisms (AI personal utility, proactive career behaviors, job insecurity, workplace anxiety) and moderating resources (organizational AI support, employee resilience), capturing the dynamic interactions between individual, organizational, and technological elements. This systemic lens allows for the simultaneous examination of positive and negative pathways, illustrating how AI technologies create both opportunities and challenges within organizational systems.

Figure 1 presents the integrative model, highlighting the dual pathways: positive reinforcement via challenge-related technostress and negative feedback via hindrance-related technostress. This framework provides a comprehensive basis for empirical testing, offering theoretical and practical insights into managing AI-induced technostress in complex organizational environments.

4. Data and Results: A Systems Perspective

4.1. Survey Design and Variable Measurement

4.1.1. Data Collection for the Socio-Technical System Model

The data for this study was collected through field surveys of employees at a large state-owned high-tech research institute in China. This organization specializes in pharmaceutical R&D and has recently implemented extensive AI systems. To ensure sample representativeness, we recruited participating companies spanning diverse high-tech-related industries. Initially, we contacted the human resources directors of target organizations to explain the research objectives and guarantee data confidentiality. Following organizational approval, we requested assistance from HR departments to randomly select employees from different divisions—including R&D, operations, and marketing—to minimize potential sampling bias. To ensure methodological rigor and minimize common method bias, the survey employed a multi-wave, dyadic design involving distinct data sources (employees and their immediate supervisors). Data collection was conducted in two specific time waves with a three-week interval.

• Stage 1 (Employee self-assessment): Employee data were collected in two waves to capture temporal dynamics.
  • Wave 1 (14 October 2024): Employees completed the initial survey measuring variables in

    Table 1. Variable Measurement Summary.

    Variable	Code	Items	Source
    AI Personal Utility	uti	4	Park, Woo, & Kim (2024); e.g., [55], “Using AI enhances my confidence in my job skills.” ($\alpha$ = 0.87)
    Job Insecurity	inse	4	Park, Woo, & Kim (2024); e.g., [55], “I worry that my job skills will be replaced by AI.” ($\alpha$ = 0.94)
    AI-related Technostress	tech (ctech/htech)	7	Ding (2021) [51]; challenge stress (4 items, $\alpha$ = 0.76), hindrance stress (3 items, $\alpha$ = 0.81)
    Organizational AI Support	support	4	Adapted from Chatterjee et al. (2021) [73]; e.g., “With company support, I have sufficient resources to learn AI.” ($\alpha$ = 0.87)
    Employee Resilience	tenacity	9	Näswall et al. (2019); e.g., [68], “I can work effectively with others to address workplace challenges.” ($\alpha$ = 0.91)
    Proactive Career Behaviors	active	3	Wu et al. (2018); e.g., [74], “I take on tasks that help advance my career.” ($\alpha$ = 0.79)
    Workplace Anxiety	anxiety	5	Parker & DeCotiis (1983); e.g., [75], “My work makes me feel anxious.” ($\alpha$ = 0.74)
    Career Growth	growth	15	Weng et al. (2010); e.g., [76], “My current job brings me closer to my career goals.” ($\alpha$ = 0.936)

    , along with demographic information.
  • Wave 2 (4 November 2024): Three weeks later, the same employees completed the second survey measuring the variables in Table 1. (Sample questionnaires for the employee self-assessment (Wave 1 and Wave 2) can be accessed at: https://www.wjx.cn/vm/OEggdMg.aspx, accessed on 14 October 2024 and https://www.wjx.cn/vm/tkxxHP2.aspx, accessed on 4 November 2024).
• Stage 2 (Supervisor evaluation): Concurrently with the second wave of employee surveys (4 November 2024), the immediate supervisors of these employees were invited to assess their subordinates’ Career Growth. This separation of data sources (employees rating stressors vs. supervisors rating outcomes), combined with the temporal separation between independent and dependent variables, was strictly implemented to minimize transient mood effects and potential common method bias. (The supervisor evaluation questionnaire can be accessed at: https://www.wjx.cn/vm/wzzEL6z.aspx, accessed on 4 November 2024).

During the data collection process, a rigorous coding system was used to match responses. Participants generated a unique non-identifiable code consisting of their initials and the last four digits of their mobile phone number (e.g., “WXM1234”). To ensure accuracy, employees with identical initials were verified in advance to prevent duplication. A total of 326 valid matched pairs were obtained from a large research institute, including 73 female participants. The mean age of respondents was 34.6 years. Attrition matched across waves was minimal due to the high stability of the workforce in this state-owned high-tech sector.

4.1.2. Variable Measurement and Analysis Strategy

Table 1 summarizes the constructs, abbreviations, item numbers, and measurement sources. All constructs were adapted from validated scales to the AI context. Items were rated on a 5-point Likert scale (1 = “strongly disagree”, 5 = “strongly agree”).

To empirically test the proposed research model, we employed Partial Least Squares Structural Equation Modeling (PLS-SEM) using the SmartPLS 4 software package. PLS-SEM was selected because this study aims to predict the variance in the target construct (career growth) and the model involves complex serial mediation pathways.

The analysis followed a two-step procedure:

• Measurement Model Assessment: We evaluated the reliability and validity of the constructs by examining outer loadings, Cronbach’s $\alpha$, Composite Reliability (CR), Average Variance Extracted (AVE), and the Heterotrait-Monotrait ratio (HTMT).
• Structural Model Assessment: We tested the hypothesized relationships (path coefficients) and specific indirect effects using a bootstrapping procedure with 5000 subsamples.

4.1.3. Descriptive and Correlation Analysis

Preliminary correlation analysis revealed that AI challenge-related technostress (ctech) is significantly positively associated with AI personal utility (uti), proactive career behaviors (active), and career growth (growth) ($p<0.01$). These findings provide initial support for the hypothesized reinforcing feedback loop between challenge-related stress and career-enhancing behaviors.

4.1.4. Measurement Model Assessment

The measurement model assessment indicated that all outer loadings exceeded 0.6. Specifically, Cronbach’s $\alpha$ values for all constructs ranged from 0.74 to 0.94, composite reliability (CR) values were consistently above 0.70, and average variance extracted (AVE) exceeded 0.50, demonstrating robust convergent validity. Discriminant validity was supported as the square root of AVE for each construct exceeded inter-construct correlations (Fornell–Larcker criterion). Furthermore, we examined the Heterotrait-Monotrait ratio (HTMT) for all construct pairs. All HTMt-Values were found to be below the strict threshold of 0.85, providing further evidence that the constructs are empirically distinct and suitable for system-level analyses.

4.2. Structural Equation Modeling and Results

Prior to examining the structural pathways, we assessed potential multicollinearity issues among the predictor variables. The Variance Inflation Factor (VIF) values for all constructs ranged from 1.21 to 2.87, which are well below the conservative threshold of 3.3. These results indicate that multicollinearity is not a significant concern in our model, ensuring the robustness of the subsequent structural analysis.

4.2.1. Direct Effects Within the System

SEM analysis confirmed that AI challenge-related technostress (ctech) positively influenced AI personal utility (uti), which in turn promoted proactive career behaviors (active) and career growth (growth), supporting hypotheses H2a and H2b. This pathway illustrates a reinforcing loop within the organizational system, where challenge-related stressors act as developmental resources.

4.2.2. Moderating Effects of Organizational AI Support

Organizational AI support (support) was found to strengthen the positive relationship between challenge-related technostress and AI personal utility (supporting H3a). Moderated mediation analysis revealed that the indirect effect on career growth via “AI personal utility → proactive career behaviors” was amplified under high AI support (supporting H3b). This indicates that organizational resources modulate the strength of reinforcing feedback loops in the socio-technical system.

4.2.3. Moderating Effects of Employee Resilience

Employee resilience (tenacity) significantly moderated the effect of challenge-related technostress on AI personal utility (supporting H4a) and the subsequent mediated pathway to career growth (supporting H4b). Resilient employees utilize system resources more effectively, sustaining positive cycles of skill acquisition and proactive behavior.

4.2.4. Robustness and Dual-Path Verification

Additional analyses confirmed that AI hindrance-related technostress (htech) heightened job insecurity and workplace anxiety, forming a balancing loop that constrains career growth. These results highlight the dual-path system dynamics of AI technostress: challenge-related stressors reinforce career-enhancing behaviors, whereas hindrance-related stressors activate negative feedback loops that limit development.

4.2.5. System-Level Implications

Overall, the empirical findings validate the proposed socio-technical model, emphasizing the interactive roles of individual, organizational, and technological resources. The results underscore that managing AI-induced technostress requires attention to system-level dynamics, ensuring that reinforcing pathways are supported while balancing mechanisms of hindrance are mitigated.

4.3. The Mechanism of AI Challenge-Related Technostress on Employees’ Career Growth (B)

4.3.1. Impact of AI Challenge-Related Technostress on Employees’ Career Growth

Direct Effect of AI Challenge-Related Technostress on Career Growth

Regression analysis reveals that AI challenge-related technostress has a significant positive effect on employees’ career growth ($\beta =0.080,p<0.001$). This suggests that employees’ perceived challenges under AI contexts can effectively stimulate their learning and development motivation, thus, enhancing career growth.

Testing the Direct Effect

After controlling for gender, age, education, and tenure, ctech still has a significant positive effect. This section examines the mechanism through which challenging AI-related technostress influences employees’ career growth. We focus on the cognitive-behavioral pathway, highlighting the mediating roles of AI personal utility and proactive career behaviors, as well as the moderating effects of organizational AI support.

Measurement Model Evaluation of the Resource Gain Chain

The outer loadings for the indicators of challenge technostress, AI personal utility, proactive career behaviors, and career growth are presented in

Table 2. Outer Loadings for the Resource Gain Chain.

Indicator	Loading (O)	t-Value	p-Value	95% CI (Bias Corrected)
ctech1 ← CTech	0.892	45.254	0.000	[0.848, 0.925] ***
ctech2 ← CTech	0.891	44.414	0.000	[0.847, 0.926] ***
ctech3 ← CTech	0.910	60.419	0.000	[0.877, 0.937] ***
ctech4 ← CTech	0.838	27.804	0.000	[0.774, 0.890] ***
uti1 ← Utility	0.886	57.266	0.000	[0.853, 0.913] ***
uti2 ← Utility	0.893	40.668	0.000	[0.845, 0.930] ***
uti3 ← Utility	0.885	57.399	0.000	[0.853, 0.913] ***
uti4 ← Utility	0.845	36.506	0.000	[0.796, 0.886] ***
acti1 ← Proactive	0.806	28.961	0.000	[0.760, 0.850] ***
acti2 ← Proactive	0.909	51.267	0.000	[0.883, 0.932] ***
acti3 ← Proactive	0.917	73.640	0.000	[0.876, 0.927] ***
grow1 ← Career Growth	0.818	34.030	0.000	[0.765, 0.859] ***
grow2 ← Career Growth	0.791	24.769	0.000	[0.722, 0.847] ***
grow3 ← Career Growth	0.864	50.006	0.000	[0.827, 0.894] ***
grow4 ← Career Growth	0.883	55.368	0.000	[0.847, 0.910] ***
grow5 ← Career Growth	0.854	43.569	0.000	[0.813, 0.890] ***
grow6 ← Career Growth	0.865	47.412	0.000	[0.825, 0.897] ***
grow7 ← Career Growth	0.867	49.323	0.000	[0.829, 0.899] ***
grow8 ← Career Growth	0.853	42.597	0.000	[0.811, 0.890] ***
grow9 ← Career Growth	0.627	11.400	0.000	[0.504, 0.717] ***
grow10 ← Career Growth	0.696	14.774	0.000	[0.588, 0.772] ***
grow11 ← Career Growth	0.669	12.101	0.000	[0.543, 0.762] ***
grow12 ← Career Growth	0.544	8.965	0.000	[0.412, 0.648] ***
grow13 ← Career Growth	0.430	5.805	0.000	[0.267, 0.558] ***
grow14 ← Career Growth	0.597	10.231	0.000	[0.469, 0.695] ***
grow15 ← Career Growth	0.414	5.415	0.000	[0.246, 0.545] ***

Note: *** p < 0.001.

. All indicators of challenge technostress (ctech1–ctech4) loaded strongly on the construct, with loadings ranging from 0.838 to 0.910, suggesting high reliability in capturing the stimulating aspects of AI-related challenges. Similarly, AI personal utility (uti1–uti4) demonstrated consistently high loadings (0.845–0.893), confirming that employees’ perceptions of AI’s usefulness were well-reflected in the measurement items. Proactive career behaviors (acti1–acti3) showed loadings above 0.80, with particularly strong results for acti2 and acti3 (>0.90), indicating robust construct validity.

Regarding the outcome variable, career growth (grow1–grow15), most items exhibited strong loadings above 0.60, particularly grow3 to grow8 (>0.85), reflecting a solid measurement foundation. While a few items (e.g., grow13 and grow15) showed lower loadings around 0.40, the majority of indicators still exceeded the threshold of 0.50, and their inclusion was theoretically justified to ensure comprehensive coverage of the construct. Collectively, the results indicate that the measurement model for the resource gain chain demonstrates adequate convergent validity, supporting the reliability of subsequent structural path testing.

The outer loadings for challenge-related technostress, AI utility, proactive career behaviors, and career growth are consistently above the recommended threshold of 0.70, with most exceeding 0.80. This indicates strong indicator reliability and convergent validity. In particular, AI utility and proactive behaviors show very high loadings, confirming their robustness as mediators in the resource gain pathway. These results suggest that employees exposed to AI-related challenges perceive clear benefits, adopt proactive behaviors, and thereby foster career growth.

4.3.2. Structural Relationships Among Latent Variables

The results provide strong evidence for the resource gain mechanism(see

Table 3. Path Coefficients of the Resource Gain Pathway.

Path	Coefficient (O)	t-Value	p-Value	95% CI (Bias Corrected)
ctech → uti	0.430	7.511	0.000	[0.319, 0.540] ***
uti → acti	0.448	8.452	0.000	[0.340, 0.547] ***
acti → grow	0.413	7.323	0.000	[0.307, 0.525] ***
tena → acti	0.254	4.314	0.000	[0.143, 0.374] ***
sup → uti	−0.011	0.183	0.855	[−0.124, 0.104] n.s.
sup × ctech → uti	0.094	1.925	0.054	[−0.002, 0.192] †
tena × uti → acti	0.054	1.114	0.265	[−0.042, 0.148] n.s.

Note: *** p < 0.001, † p ≈ 0.05, n.s. = not significant.

). Challenge-related technostress (ctech) exerts a significant positive effect on AI personal utility (uti), which in turn strongly enhances proactive career behaviors (acti). These proactive behaviors significantly contribute to career growth (grow), suggesting that employees can transform technological stress into developmental opportunities when they perceive AI as useful. Moreover, organizational AI support (sup) does not have a direct effect on utility, but its interaction with challenge stress is marginally significant, indicating a boundary condition where support may amplify the positive role of challenge stress. Resilience (tena) significantly promotes proactive behaviors directly, although its moderation effect on the “utility → proactive behavior” link is not significant. Together, these findings confirm that challenge-related stress operates through a resource gain pathway, enabling employees to actively leverage AI for career advancement.

In the structural model, the hypothesized relationships among variables were empirically supported. First, Proactive Career Behaviors exert a significant positive effect on Career Growth ($\beta =0.413$, $p<0.001$), suggesting that employees who engage more actively in career behaviors experience higher levels of growth. Conversely, Workplace Anxiety shows a significant negative impact on career growth ($\beta =-0.253$, $p<0.001$), indicating that anxiety hinders employees’ development.

Furthermore, AI Personal Utility significantly enhances proactive career behaviors ($\beta =0.448$, $p<0.001$), and one of its antecedents, Challenge Technostress, exerts a significant positive influence on AI personal utility ($\beta =0.430$, $p<0.001$). This implies that when employees perceive AI-related technological challenges as opportunities for growth, they are more likely to effectively utilize AI tools, which in turn fosters proactive career behaviors.

In terms of psychological mechanisms, Hindrance Technostress significantly increases Job Insecurity ($\beta =0.575$, $p<0.001$), whereas Organizational AI Support significantly alleviates job insecurity ($\beta =-0.460$, $p<0.001$). However, organizational AI support does not significantly influence AI personal utility ($\beta =-0.011$, $p=0.855$).

Moreover, Employee Resilience plays a pivotal role in the model. On the one hand, it directly enhances proactive career behaviors ($\beta =0.254$, $p<0.001$); on the other hand, it significantly reduces workplace anxiety ($\beta =-0.149$, $p<0.01$). Notably, the interaction between employee resilience and job insecurity further intensifies workplace anxiety ($\beta =0.117$, $p<0.05$), suggesting that in environments with high job insecurity, even highly resilient employees may experience additional psychological strain.

Taken together, these findings reveal a clear causal chain: organizational and technological conditions (e.g., challenge and hindrance technostress, organizational AI support) shape AI personal utility, job insecurity, and workplace anxiety, which in turn affect proactive career behaviors and ultimately career growth. This implies that managers aiming to foster employees’ career growth should simultaneously enhance the personal utility of AI tools, mitigate hindrance technostress, strengthen organizational AI support, and pay attention to employees’ resilience and workplace anxiety levels.

4.3.3. Testing Specific Indirect Effects

The analysis of the specific indirect effects provides further insights into the mediating mechanisms underlying the relationships among the constructs (see

Table 4. Specific Indirect Effects of the Model.

Indirect Path	Estimate (O)	t-Value	p-Value	Significance
AI Personal Utility → Proactive Career Behaviors → Career Growth	0.185	6.518	0.000	***
Challenge Technostress → AI Personal Utility → Proactive Career Behaviors → Career Growth	0.080	4.771	0.000	***
Employee Resilience → Proactive Career Behaviors → Career Growth	0.105	3.340	0.001	**
Employee Resilience → Workplace Anxiety → Career Growth	0.038	2.050	0.040	*
Employee Resilience × Job Insecurity → Workplace Anxiety → Career Growth	−0.030	1.988	0.047	*
Challenge Technostress → AI Personal Utility → Proactive Career Behaviors	0.193	6.691	0.000	***
Job Insecurity → Workplace Anxiety → Career Growth	−0.028	1.811	0.070	†
Hindrance Technostress → Job Insecurity → Workplace Anxiety → Career Growth	−0.016	1.779	0.075	†
Organizational AI Support × Challenge Technostress → AI Personal Utility → Proactive Career Behaviors → Career Growth	0.017	1.788	0.074	†
Organizational AI Support → Job Insecurity → Workplace Anxiety	−0.051	1.696	0.090	n.s.
Organizational AI Support → Job Insecurity → Workplace Anxiety → Career Growth	0.013	1.693	0.091	n.s.
Organizational AI Support → AI Personal Utility → Proactive Career Behaviors → Career Growth	−0.002	0.182	0.855	n.s.
Organizational AI Support → AI Personal Utility → Proactive Career Behaviors	−0.005	0.185	0.854	n.s.
Employee Resilience × AI Personal Utility → Proactive Career Behaviors → Career Growth	0.022	1.093	0.274	n.s.
Organizational AI Support × Hindrance Technostress → Job Insecurity → Workplace Anxiety	−0.001	0.121	0.904	n.s.
Organizational AI Support × Hindrance Technostress → Job Insecurity → Workplace Anxiety → Career Growth	0.000	0.119	0.905	n.s.
Hindrance Technostress → Job Insecurity → Workplace Anxiety	0.064	1.791	0.073	†

Notes: *** p < 0.001, ** p < 0.01, * p < 0.05, † p < 0.10, n.s. = not significant.

). The results reveal that AI Personal Utility significantly contributes to Career Growth through Proactive Career Behaviors. Specifically, the path “AI Personal Utility → Proactive Career Behaviors → Career Growth” demonstrates a strong positive indirect effect ($\beta =0.185$, $t=6.518$, $p<0.001$), suggesting that employees’ perception of AI usefulness promotes proactive behaviors, which in turn enhance their career advancement. A similar but smaller pathway, “Challenge Technostress → AI Personal Utility → Proactive Career Behaviors → Career Growth,” also shows significance ($\beta =0.080$, $t=4.771$, $p<0.001$), indicating that challenge-oriented technostress can foster positive outcomes when employees leverage AI in their work.

Employee Resilience emerges as another key driver of career outcomes. Two significant mediation paths are observed: “Employee Resilience → Proactive Career Behaviors → Career Growth” ($\beta =0.105$, $t=3.340$, $p=0.001$) and “Employee Resilience → Workplace Anxiety → Career Growth” ($\beta =0.038$, $t=2.050$, $p=0.040$). These findings suggest that resilient employees are more likely to engage proactively with their careers and manage workplace anxiety effectively, thereby achieving higher levels of career growth.

Moreover, the interaction between Employee Resilience and Job Insecurity exhibits a significant negative indirect effect on career outcomes. The path “Employee Resilience × Job Insecurity → Workplace Anxiety → Career Growth” ($\beta =-0.030$, $t=1.988$, $p<0.05$) indicates that even resilient employees may experience detrimental effects on career development when facing heightened job insecurity, primarily through increased workplace anxiety.

In contrast, several paths involving Organizational AI Support and Hindrance Technostress do not yield significant indirect effects (e.g., “Organizational AI Support → Job Insecurity → Workplace Anxiety” and “Organizational AI Support → AI Personal Utility → Proactive Career Behaviors”), implying that organizational support alone may not be sufficient to buffer the negative consequences of hindrance stressors or job insecurity.

Overall, these findings highlight that AI Personal Utility and Employee Resilience are the most critical enablers of Career Growth, primarily through their impact on Proactive Career Behaviors and the reduction of Workplace Anxiety. Conversely, Job Insecurity and Hindrance Technostress tend to function as risk factors, with limited or negative indirect contributions to career outcomes.

Chain Mediation Pathway

The first chain mediation pathway can be summarized as follows: Challenging Technostress → AI Personal Utility → Proactive Career Behaviors → Career Growth (see in

Table 5. Specific Indirect Effects of the Resource Gain Pathway.

Indirect Path	Coefficient (O)	t-Value	p-Value	95% CI (Bias Corrected)
ctech → uti → acti	0.193	6.691	0.000	[0.136, 0.250] ***
ctech → uti → acti → grow	0.080	4.771	0.000	[0.049, 0.115] ***
uti → acti → grow	0.185	6.518	0.000	[0.131, 0.242] ***
tena → acti → grow	0.105	3.340	0.001	[0.052, 0.174] **
sup × ctech → uti → acti	0.042	1.708	0.088	[−0.001, 0.097] †
sup × ctech → uti → acti → grow	0.017	1.788	0.074	[0.000, 0.038] †
tena × uti → acti → grow	0.022	1.093	0.274	[−0.018, 0.062] n.s.
sup → uti → acti	−0.005	0.185	0.854	[−0.054, 0.048] n.s.
sup → uti → acti → grow	−0.002	0.182	0.855	[−0.024, 0.019] n.s.

Note: *** p < 0.001, ** p < 0.01, † p ≈ 0.05, n.s. = not significant.

). Empirical results show that challenging technostress significantly and positively predicts AI personal utility (red “+”), indicating that when employees face technology-related challenges that promote growth and capability enhancement, they are more likely to recognize the value and usefulness of AI tools. In turn, AI personal utility significantly and positively predicts proactive career behaviors (red “+”), suggesting that employees who perceive AI as helpful are more willing to take active steps to shape their career paths. Finally, proactive career behaviors significantly and positively predict career growth (red “+”), highlighting the critical role of proactive engagement in career development. The overall chain mediation effect is significant, confirming that AI personal utility and proactive career behaviors jointly bridge the effect of challenging technostress on career growth. This pathway represents a positive and effective channel for employees’ professional development.

Moderating Role of Organizational AI Support

Further analysis indicates that organizational AI support positively moderates the relationship between challenging technostress and AI personal utility. When organizations provide higher levels of AI support, such as training, resources, and institutional encouragement, the positive effect of challenging technostress on employees’ perceived AI utility is strengthened. This implies that organizational resources enhance employees’ ability to reframe technological challenges as opportunities, thereby promoting proactive behaviors and facilitating career growth.

Summary of Mechanism

In summary, challenging AI-related technostress promotes career growth through a cognitive-behavioral chain, in which AI personal utility and proactive career behaviors play pivotal mediating roles. Organizational AI support serves as an important boundary condition that amplifies the positive effect of challenging technostress on perceived AI utility. These findings provide empirical evidence for the dual-resource gain mechanism, demonstrating how both individual cognitive appraisal and organizational support jointly contribute to leveraging AI-related challenges for career development.

We employed the bootstrapping approach (with 5000 resamples) within SmartPLS to validate the hypothesized serial mediation model. The analysis confirmed that the chain mediation pathway—“Hindrance-related technostress → Job insecurity → Workplace anxiety → Career growth”—was supported.

Specifically, disruptive technostress significantly and positively predicted job insecurity, indicating that when employees face technological hindrances, ambiguity, and lack of control, they are more likely to perceive instability in their jobs. Furthermore, job insecurity significantly and positively predicted workplace anxiety, suggesting that cognitive insecurity is transmitted to the emotional domain, resulting in heightened anxiety and tension. In turn, workplace anxiety significantly and negatively predicted career growth, demonstrating that anxiety consumes individual resources, weakens professional efficacy and proactive engagement and, thus, suppresses career development. Overall, the significant chain mediation effect indicates that negative technostress impacts career outcomes through the cognitive–emotional mechanism of “job insecurity → workplace anxiety,” forming a typical transmission chain of occupational predicament.

In addition, the moderation test revealed that employee resilience exerts a significant positive moderating effect on the “job insecurity → workplace anxiety” path. Specifically, under conditions of high job insecurity, employees with stronger psychological resilience were more likely to experience workplace anxiety. This may be because highly resilient individuals continue to invest cognitive and emotional resources even in the face of intense uncertainty, which in turn leads to deeper emotional reactions.

In summary, challenging technostress can promote career growth through a cognitive-behavioral chain mechanism, consisting of AI personal utility and proactive career behaviors. Organizational AI support acts as a crucial boundary condition that amplifies the positive effect of challenging technostress on perceived AI utility, enhancing the overall pathway toward career development. These findings provide empirical support for the dual-resource gain mechanism, highlighting both cognitive appraisal and behavioral engagement in leveraging AI-related challenges.

4.4. The Mechanism of AI Hindrance-Related Technostress on Employees’ Career Growth (C)

4.4.1. Impact of AI Hindrance-Related Technostress on Employees’ Career Growth

Direct Effect of AI Hindrance-Related Technostress on Career Growth

Regression analysis indicates that AI hindrance-related technostress exerts a negative but marginally significant effect on employees’ career growth ($\beta =-0.016,p=0.075$). This result suggests that when employees perceive AI as a source of obstacles, overload, or role threats, their career development tends to be hindered rather than facilitated.

Testing the Direct Effect

After controlling for gender, age, education, and tenure, htech still exhibits a significant negative effect. This section therefore explores the mediating and moderating mechanisms through which AI hindrance-related technostress undermines employees’ career growth, with a focus on the sequential mediation of job insecurity and workplace anxiety, and the moderating roles of organizational AI support and employee resilience.

Measurement Model Evaluation of the Resource Loss Chain

Table 6. Outer Loadings for the Resource Loss Chain.

Indicator	Loading (O)	t-Value	p-Value	95% CI (Bias Corrected)
htech1 ← HTech	0.946	78.065	0.000	[0.920, 0.967] ***
htech2 ← HTech	0.968	134.635	0.000	[0.952, 0.980] ***
htech3 ← HTech	0.956	107.239	0.000	[0.936, 0.971] ***
inse1 ← Job Insecurity	0.887	63.230	0.000	[0.858, 0.912] ***
inse2 ← Job Insecurity	0.893	51.935	0.000	[0.856, 0.923] ***
inse3 ← Job Insecurity	0.905	66.279	0.000	[0.875, 0.928] ***
inse4 ← Job Insecurity	0.782	23.990	0.000	[0.710, 0.837] ***
anxie1 ← Anxiety	0.841	32.724	0.000	[0.786, 0.905] ***
anxie2 ← Anxiety	0.865	27.996	0.000	[0.860, 0.883] ***
anxie3 ← Anxiety	0.908	73.238	0.000	[0.883, 0.932] ***
anxie4 ← Anxiety	0.907	68.957	0.000	[0.876, 0.927] ***
anxie5 ← Anxiety	0.501	6.447	0.000	[0.323, 0.630] ***
grow1 ← Career Growth	0.818	34.030	0.000	[0.765, 0.859] ***
grow2 ← Career Growth	0.791	24.769	0.000	[0.722, 0.847] ***
grow3 ← Career Growth	0.864	50.006	0.000	[0.827, 0.894] ***
grow4 ← Career Growth	0.883	55.368	0.000	[0.847, 0.910] ***
grow5 ← Career Growth	0.854	43.569	0.000	[0.813, 0.890] ***
grow6 ← Career Growth	0.865	47.412	0.000	[0.825, 0.897] ***
grow7 ← Career Growth	0.867	49.323	0.000	[0.829, 0.899] ***
grow8 ← Career Growth	0.853	42.597	0.000	[0.811, 0.890] ***
grow9 ← Career Growth	0.627	11.400	0.000	[0.504, 0.717] ***
grow10 ← Career Growth	0.696	14.774	0.000	[0.588, 0.772] ***
grow11 ← Career Growth	0.669	12.101	0.000	[0.543, 0.762] ***
grow12 ← Career Growth	0.544	8.965	0.000	[0.412, 0.648] ***
grow13 ← Career Growth	0.430	5.805	0.000	[0.267, 0.558] ***
grow14 ← Career Growth	0.597	10.231	0.000	[0.469, 0.695] ***
grow15 ← Career Growth	0.414	5.415	0.000	[0.246, 0.545] ***

Note: *** p < 0.001.

reports the outer loadings for hindrance technostress, job insecurity, workplace anxiety, and career growth. The hindrance technostress construct (htech1–htech3) showed extremely high loadings, ranging from 0.946 to 0.968, indicating excellent reliability in capturing employees’ perceptions of AI-related technological hindrances. Job insecurity (inse1–inse4) also displayed strong loadings (0.782–0.905), suggesting that the items effectively represented the latent construct.

For workplace anxiety, four of the five indicators (anxie1–anxie4) loaded above 0.84, reflecting high measurement validity. However, anxie5 exhibited a relatively low loading of 0.501, which—though still above the minimum acceptable threshold of 0.50—suggests weaker representation and should be interpreted with caution. The career growth indicators (grow1–grow15) followed a similar pattern as in the resource gain chain, with the majority of loadings exceeding 0.60 and several exceeding 0.85, ensuring comprehensive construct representation.

Overall, the results demonstrate that the measurement model for the resource loss chain exhibits strong convergent validity, with only minor concerns regarding one anxiety indicator. These findings provide confidence in the robustness of the measurement model and lend credibility to the subsequent analysis of indirect effects through job insecurity and workplace anxiety.

For hindrance-related technostress, job insecurity, workplace anxiety, and career growth, most loadings also exceed 0.80, although one anxiety indicator (Anxie5) shows a relatively weak loading (0.501). Despite this, the measurement quality remains acceptable overall. The strong loadings for hindrance stress and job insecurity confirm the reliability of the resource loss constructs. This supports the view that AI-related hindrance stressors primarily deplete resources through insecurity and anxiety, ultimately constraining career growth.

4.4.2. Effects Among Latent Variables

The structural model results support the hypothesized resource-depletion pathway (see in

Table 7. Significance of Path Coefficients: Hindrance-Related Technostress Model.

Path	Coefficient (O)	t-Value	p-Value	95% CI (Bias Corrected)
htech → inse	0.575	12.342	0.000	[0.481, 0.662] ***
inse → anxie	0.111	1.825	0.068	[−0.009, 0.229] †
anxie → grow	−0.253	5.243	0.000	[−0.341, −0.147] ***
sup → inse	−0.460	7.509	0.000	[−0.569, −0.326] ***
resi → anxie	−0.149	2.495	0.013	[−0.257, −0.025] **
resi → grow	0.087	2.140	0.033	[0.008, 0.166] *
sup × htech → inse	−0.007	0.134	0.894	[−0.099, 0.108] n.s.
resi × inse → anxie	0.117	2.159	0.031	[0.010, 0.224] *

Note: *** p < 0.001, ** p < 0.01, * p < 0.05, † p ≈ 0.05, n.s. = not significant.

). First, AI Hindrance-Related Technostress strongly increases Job Insecurity ($\beta =0.575$, $p<0.001$). In turn, Job Insecurity is positively associated with Workplace Anxiety ($\beta =0.111$, $p\approx 0.07$), although the effect is marginal. Finally, Workplace Anxiety exerts a significant negative effect on Career Growth ($\beta =-0.253$, $p<0.001$).

With respect to contextual moderators, Organizational AI Support significantly reduces job insecurity ($\beta =-0.460$, $p<0.001$), indicating that supportive training and resource allocation mitigate insecurity perceptions. Similarly, Employee Resilience alleviates workplace anxiety ($\beta =-0.149$, $p<0.01$) and directly enhances career growth ($\beta =0.087$, $p<0.05$). Interestingly, the interaction effect between employee resilience and job insecurity is positive ($\beta =0.117$, $p<0.05$). While counter-intuitive, this reveals a potential ’dark side’ of resilience within the socio-technical framework. Drawing on the Conservation of Resources (COR) theory, highly resilient employees typically invest heavily in their professional identity and active coping strategies. Under conditions of high AI-induced job insecurity, these individuals may experience greater cognitive load and emotional strain because they have more psychological resources at stake (i.e., ’sunk costs’ in their career path) compared to less engaged employees. Thus, in high-threat environments, the very effort to maintain resilience may paradoxically fuel heightened anxiety.

The findings highlight the resource depletion mechanism. Hindrance-related technostress (htech) significantly heightens job insecurity (inse), although insecurity does not significantly increase workplace anxiety (anxie). Nevertheless, workplace anxiety exerts a strong and negative effect on career growth, confirming its detrimental role in employees’ development. Importantly, organizational AI support (sup) mitigates job insecurity significantly, serving as a protective organizational resource. Furthermore, resilience (tena) directly reduces anxiety and moderates the insecurity–anxiety relationship positively, indicating that resilient employees are better equipped to buffer the psychological toll of insecurity. These results suggest that when employees perceive AI as a threat rather than an opportunity, resources are depleted, leading to heightened anxiety and undermined career growth, unless counteracted by organizational support and individual resilience.

Furthermore, our results clarify the primacy of the ’social’ subsystem over the ’technical’ subsystem in determining career outcomes. The finding that hindrance-related technostress exerts only a marginal direct effect on career growth ($p=0.075$), while its indirect impact via job insecurity is significant, supports a ’full mediation’ mechanism. This suggests that AI technologies—even when complex or intrusive—do not directly impede career development. Instead, their negative impact is almost entirely contingent on whether they trigger a psychological sense of threat. This validates our socio-technical premise that individual psychological appraisal acts as the decisive filter, effectively ’gating’ the impact of technological stressors.

4.4.3. Analysis of Specific Indirect Effects

Chain Mediation Pathway

The analysis of specific indirect effects reveals the existence of a partial chain mediation (see in

Table 8. Specific Indirect Effects of Hindrance-Related Technostress Model.

Indirect Path	Estimate (O)	t-Value	p-Value	Significance
Job Insecurity → Workplace Anxiety → Career Growth	−0.028	1.811	0.070	†
Hindrance Technostress → Job Insecurity → Workplace Anxiety → Career Growth	−0.016	1.779	0.075	†
Organizational AI Support → Job Insecurity → Workplace Anxiety → Career Growth	0.013	1.693	0.091	n.s.
Employee Resilience → Workplace Anxiety → Career Growth	0.038	2.050	0.040	*
Employee Resilience × Job Insecurity → Workplace Anxiety → Career Growth	−0.030	1.988	0.047	*

Notes: * p < 0.05, † p < 0.10, n.s. = not significant.

). The path “Hindrance Technostress → Job Insecurity → Workplace Anxiety → Career Growth” is negative and marginally significant ($\beta =-0.016$, $t=1.779$, $p\approx 0.075$), consistent with the hypothesized resource-depletion process. Moreover, Employee Resilience demonstrates a protective role: the indirect effect “Resilience → Workplace Anxiety → Career Growth” is significant and positive ($\beta =0.038$, $p<0.05$), indicating that resilient employees can partially counteract the detrimental impact of anxiety on growth. However, the interaction path “Resilience × Job Insecurity → Workplace Anxiety → Career Growth” is negative ($\beta =-0.030$, $p<0.05$), suggesting that resilience may have boundary conditions under high-insecurity contexts.

Summary of Mechanism

Overall, these results confirm that AI hindrance-related technostress undermines career growth primarily through a chain of job insecurity and workplace anxiety. While organizational AI support and employee resilience buffer some of the negative effects, they cannot fully offset the resource-depletion pathway. This underscores the importance of managing AI hindrance stressors at both organizational and individual levels. It is worth noting that while our data collection focused on the individual level, this aligns with the micro-foundational perspective of organizational systems. We posit that the effectiveness of a system approach is inherently reflected in and enacted through individual behaviors and perceptions. Thus, individual-level data serves as a valid proxy for evaluating the operational dynamics of the organizational system.

5. Conclusions: A Systems Perspective on AI-Related Technostress

This study adopts a systems-oriented, dual-pathway perspective to examine how AI-related technostress shapes employees’ career growth. Integrating the Job Demands–Resources (JD–R) framework with the Transactional Theory of Stress, we identify two interacting subsystems.

• Summary of Dual-Path Mechanisms

The resource-gain loop operates as a self-reinforcing system: Challenge-Related Technostress positively influences AI Personal Utility, triggering Proactive Career Behaviors that accelerate Career Growth. Organizational AI Support amplifies this positive spiral. In contrast, the resource-depletion loop sees Hindrance-Related Technostress elevating Job Insecurity, which propagates to Workplace Anxiety and ultimately stifles growth. Crucially, our findings regarding Employee Resilience reveal a nuanced dynamic: rather than functioning purely as a buffer, high resilience was found to amplify the impact of job insecurity on anxiety. This suggests that highly engaged employees may experience greater psychological strain when their professional continuity is threatened.

• Theoretical Contributions

This study moves beyond general dual-pathway rhetoric to offer three concrete theoretical advancements. First, we disentangle the distinct micro-foundations of adaptation. Unlike prior studies treating outcomes monolithically, we identify AI Personal Utility as the specific cognitive driver of the gain loop, and Workplace Anxiety as the distinct emotional barrier in the loss loop. This distinction clarifies the “black box” of socio-technical adaptation [77]. Second, we extend the JD–R model by uncovering a “dark side” of personal resources. Deviating from the traditional buffering hypothesis, the positive interaction effect demonstrates that resilience can intensify the anxiety triggered by insecurity. This challenges the assumption that personal resources always mitigate stress. Third, our empirical examination of the serial mediation path (Hindrance → Insecurity → Anxiety) validates the sequential nature of resource depletion. This enriches the systemic view by mapping exactly how negative feedback loops are sustained through specific cognitive-affective transitions.

• Practical Implications

Managers should design interventions that simultaneously enhance reinforcing loops and mitigate depleting loops: (1) Reframe AI demands as developmental challenges through training and institutional support. Such training initiatives should consider diverse educational perspectives and learning contexts to be effective [78]. (2) Enhance employees’ perception of AI utility to promote proactive career behaviors. (3) Monitor and reduce job insecurity to prevent the activation of resource-depletion loops. (4) Cultivate resilience while managing its potential downside; recognize that highly resilient employees may require targeted reassurance under high insecurity. These strategies align with broader insights into digital adoption, suggesting that overcoming challenges relies on robust socio-technical integration [79].

Managerial Takeaway: Effective AI management requires a balanced approach: organizations must actively promote the resource-gain loop by enhancing utility perception, while simultaneously disrupting the resource-depletion loop by reducing insecurity. Crucially, managers must remain vigilant regarding high-resilience employees, providing them with reassurance rather than assuming they are immune to stress.

• Concluding Remarks

AI technostress is a dual-edged, dynamic phenomenon. When managed within a supportive system, challenge-related stress reinforces adaptability and growth. Conversely, unmanaged hindrance stress triggers resource-depletion loops that erode well-being. This systems perspective underscores the necessity of synchronized individual, technological, and organizational interventions to navigate AI-driven workplace transformations successfully. While our study captures these dynamics using cross-sectional data, future research could employ advanced data-driven approaches designed for modeling complex, evolving dynamic systems [80,81,82] to track these feedback loops over time.