Abstract
Purpose: Organisations increasingly invest in collaborative artificial intelligence (AI) literacy to improve employee performance, yet many simultaneously face rising technostress. This creates a critical managerial dilemma: when do AI capability investments generate performance gains, and when do they erode the perceived fit between employees and AI-enabled work systems? This study examines how collaborative AI literacy translates into employee performance under varying levels of technostress.
Design/methodology/approach: Using a two-wave, multi-source survey of 403 employee–supervisor dyads. This study tests a conditional process model in which task–technology fit (TTF) explains the performance effects of collaborative AI literacy, while technostress acts as a boundary condition. Structural equation modelling with bootstrapping is employed to assess direct, indirect and interaction effects.
Findings/results: Results show that collaborative AI literacy improves both technology-enabled and creative performance primarily by strengthening TTF. However, technostress significantly weakens this relationship. Under high technostress, increases in AI literacy no longer enhance TTF and may even undermine it, resulting in diminished or reversed performance benefits. These findings reveal that AI literacy investments yield positive returns only when technostress is effectively managed.
Practical implications: Managers should recognise that AI literacy investments are contingent upon the prevailing stress environment. Organisations are advised to redesign training programmes to incorporate technostress resilience, prioritise task-technology alignment during AI system implementation, and treat technostress mitigation as a managerial performance metric. Such an integrated strategy ensures that AI capability development translates into sustainable improvements in operational efficiency and employees’ performance.
Originality/value: This study advances decision-oriented research by identifying a critical trade-off in AI capability investments: more AI literacy is not universally beneficial. By demonstrating that technostress can disrupt – and potentially reverse – the pathway from AI capability to performance, the study provides actionable guidance for managers on when to scale AI training and when to prioritise stress reduction and system simplification.
Keywords: collaborative AI literacy; task-technology fit; technostress; employee performance; task-technology fit theory.
Introduction
The global integration of artificial intelligence (AI) into the workplace is accelerating at an unprecedented pace, fundamentally reshaping the nature of work, productivity and human–machine interaction. According to a recent McKinsey report, AI adoption has more than doubled since 2017, with 55% of organisations now deploying AI in at least one business function – a trend that surged in the wake of the pandemic and continues to grow as generative AI enters mainstream use (McKinsey, 2023). However, this rapid digital transformation poses a critical managerial dilemma. While organisations aim to enhance performance through AI capability investments, employees increasingly experience technostress – overload, uncertainty and complexity induced by advanced systems (Ragu-Nathan et al., 2008) – raising the question: under what conditions do AI literacy programmes deliver real performance benefits, and when might they inadvertently reduce efficiency or task–technology alignment?
Employees require more than technical skills to work effectively with AI; they need collaborative AI literacy, defined as the capability to critically assess, ethically use and integrate AI systems in group workflows (Sidra & Mason, 2025). Although collaborative AI literacy is expected to improve performance, its benefits may only materialise when employees perceive AI tools as relevant to their work (Przegalinska et al., 2025). Task–technology fit (TTF) theory posits that performance is maximised when technology aligns with work requirements and user capabilities (Goodhue & Thompson, 1995). Therefore, collaborative AI literacy can enhance performance indirectly by strengthening perceived TTF. However, this pathway is vulnerable to disruption from technostress, particularly techno-complexity – the perception that technology is overly complex and difficult to utilise (Tarafdar et al., 2007). In AI-intensive contexts, where systems evolve rapidly, techno-complexity is both a measurable and practically significant stressor, as confirmed in prior studies that emphasise clarity and parsimony in assessing stress dimensions (Ragu-Nathan et al., 2008; Wang et al., 2023). From a managerial perspective, high techno-complexity can diminish the returns on AI training investments if not actively managed.
Collaborative AI literacy is conceptually distinct from constructs such as digital literacy, computer self-efficacy or AI self-efficacy (Sidra & Mason, 2025). While digital literacy emphasises basic proficiency and AI self-efficacy focuses on confidence in tool use, collaborative AI literacy captures the ability to critically evaluate AI outputs, integrate AI recommendations into shared workflows, and engage with AI as socio-technical collaborators. These coordination- and ethics-oriented competencies are not fully addressed by existing constructs, making collaborative AI literacy a unique and contextually vital capability. Yet, significant gaps remain for managerial application. Firstly, the mediating process through which collaborative AI literacy enhances performance – specifically via TTF – is not well understood. Secondly, although technostress is recognised as a constraint, its role as a boundary condition that can weaken the literacy–fit link has received limited empirical attention. Thirdly, methodologically robust, multi-source designs are needed to provide actionable insights into AI investments in high-tech workplaces. While our previous research identified what prevents AI adoption, the present study investigates what enables AI to improve performance once adopted (Khan et al., 2026; Jiang et al., 2026; Nawaz et al., 2025). Addressing these gaps is essential for organisations to maximise returns on AI capability investments while safeguarding employee flexibility, engagement and innovation.
A pattern has emerged across our studies: success is never automatic. Export growth requires accurate forecasting access requires barrier removal, employee performance requires low technostress (Khan et al., 2025a; 2025b; Nawaz & Liu, 2025; Nawaz et al., 2024). Given that, this study addresses the following managerial questions: Under what conditions does collaborative AI literacy improve technology-enabled and creative performance? Does TTF mediate this relationship, and is this mediation contingent on technostress – specifically techno-complexity – such that high stress reduces the effectiveness of AI training programmes? To answer these questions, we test a first-stage moderated mediation model using data from China’s services industry – a setting marked by rapid AI adoption and high technological demands (conceptual framework presented in Figure 1). The article proceeds as follows. Firstly, we review the literature and develop hypotheses grounded in TTF theory and the transactional approach to stress. Secondly, we describe the two-wave, multi-source research design and analytical methods. Thirdly, we present the results, followed by a discussion of theoretical contributions, practical implications, limitations and future research avenues. In doing so, the study offers managers an evidence-driven framework for optimising AI-driven performance while proactively managing human factors in technology integration.
Literature review and hypotheses development
Theoretical foundation
The present study is grounded in TTF theory (Goodhue & Thompson, 1995), which provides a managerial perspective on how collaborative AI literacy can translate into improved performance. Developed in the 1990s as a response to user-acceptance models like Technology Acceptance Model (TAM), TTF emphasises that performance gains are not automatic; they depend on users’ perception of fit between tasks and technology. Over the past three decades, TTF has been applied to various technologies, including ERP systems and mobile applications (Cheng, 2020), consistently showing that higher perceived fit enhances utilisation and performance. Although originally applied to static systems, TTF is highly relevant to dynamic AI environments, where flexible intelligent agents require employees to possess collaborative AI literacy. From a managerial standpoint, literacy shapes perceived fit, making TTF the key mediator through which AI investments influence performance outcomes.
To define boundary conditions, we integrate the transactional theory of stress (Lazarus & Folkman, 1984), which focuses on cognitive appraisal rather than stimulus-response models, highlighting that stress arises when perceived demands exceed coping resources (Coyne & Holroyd, 1982). In organisations, this perspective explains how workplace stressors affect attitudes, behaviours and well-being. Technostress, especially techno-complexity, is a critical factor in AI-intensive work, posing a managerial risk by potentially weakening the effectiveness of AI literacy initiatives (Azpíroz-Dorronsoro et al., 2024). Even highly literate employees may perceive complex AI interactions as overwhelming under high technostress, reducing TTF and performance gains. By combining TTF and transactional stress theory, this study offers a decision-oriented, contingent framework that helps managers understand how collaborative AI literacy improves performance and under which stress conditions these investments may be less effective or even counterproductive. This reframing emphasises actionable managerial insights over purely academic mediation logic.
Collaborative artificial intelligence literacy and task-technology fit
Modern workplaces enhanced by AI are becoming increasingly interactive, as employees work with intelligent systems as partners rather than just operators. This shift requires a new skill, described as collaborative AI literacy – the ability to evaluate, use and understand the ethical implications of AI in individual task settings. Unlike conventional digital literacy, collaborative AI literacy predicts socio-technical integration, enabling employees to make sense of AI outputs, integrate recommendations into common workflows, and coordinate with human and artificial actors (Przegalinska et al., 2025). From a managerial viewpoint, the extent to which employees leverage AI systems influences their perception of alignment between technology and tasks (Matsepe & Van der Lingen, 2022; Pranteddu et al., 2024).
Previous studies have found that competencies such as computer self-efficacy and data literacy are correlated with higher perceptions of TTF (Klopping & McKinney, 2004; Janssen et al., 2017). However, these studies focus on individual, tool-based skills rather than the collaborative, interactive capabilities required in complex human–AI work. Employees with higher collaborative AI literacy develop a more nuanced understanding of AI capabilities, anticipate how AI can support workflows, and adjust their behaviours to optimise task–technology alignment (Przegalinska et al., 2025). This knowledge helps them view AI as a logical extension of their work, enhancing perceived fit (Chin et al., 2025). In contrast, employees with lower literacy may perceive AI tools as opaque or mismatched, weakening TTF. Grounded in TTF theory (Goodhue & Thompson, 1995), collaborative AI literacy serves as a key capability for improving perceived alignment between technology and work; thus, we reached to the below given hypothesis:
H1: Collaborative AI literacy is positively related to task–technology fit.
Task–technology fit as a mediator
Based on the relationship between collaborative AI literacy and TTF, we argue that TTF serves as a key mechanism through which collaborative AI literacy translates into improved job performance. Task–technology fit theory suggests that employees who perceive strong alignment between their tasks and the technologies provided are better positioned to apply those technologies effectively, enhancing performance outcomes (Chakraborty et al., 2025). Performance in modern, technology-driven work can be conceptualised in two dimensions (Yang & Sun, 2025): technology-enabled performance, reflecting efficiency and effectiveness in executing technology-supported tasks, and creative performance, referring to the generation and implementation of novel, useful ideas.
Empirical evidence supports a strong positive relationship between TTF and performance. Studies across various technological contexts show that greater perceived fit improves task efficiency, productivity and overall job performance (Aguirre-Urreta & Marakas, 2025; Khan et al., 2025a). When technology is perceived as a good fit, cognitive load is reduced, freeing mental resources for more complex and creative work (Jia et al., 2024). In knowledge-based workplaces, high TTF has been linked not only to enhanced routine performance but also to increased innovative behaviour and creative output (Lee et al., 2025). Thus, collaborative AI literacy does not directly boost performance in isolation; its primary effect is enhancing the perception of fit. This strengthened TTF subsequently drives both technology-enabled performance and creative performance (Zhang et al., 2025a, 2025b). From a managerial perspective, TTF represents the critical pathway that converts AI literacy investments into measurable performance benefits; thus, we reached the following hypotheses:
H2a: Task–technology fit mediates the positive relationship between collaborative AI literacy and technology-enabled performance.
H2b: Task–technology fit mediates the positive relationship between collaborative AI literacy and creative performance.
Moderating role of technostress
Although collaborative AI literacy is expected to enhance TTF, this positive effect is likely contingent on contextual factors. We propose that technostress, specifically the techno-complexity dimension, serves as a critical boundary condition that may weaken this relationship. Techno-complexity refers to the stress experienced when employees perceive technology as excessively complex, difficult to learn or challenging to operate efficiently (Ragu-Nathan et al., 2008). This form of stress is particularly relevant in AI-enhanced workplaces (Kumar et al., 2024), where systems are advanced, constantly evolving and can significantly affect how employees evaluate and leverage their AI skills (Wang et al., 2025). This argument is grounded in the Transactional Theory of Stress (Lazarus & Folkman, 1984), which posits that stress arises when individuals perceive environmental demands as exceeding their coping capacity.
Collaborative AI literacy is a key personal resource for managing these demands (Sidra & Mason, 2025); however, when employees face high perceived complexity, cognitive and emotional resources may be depleted, leading to strain (Jankelova et al., 2025; Li & Wang, 2021). Even highly literate employees may feel overwhelmed, frustrated or cognitively exhausted under high technostress (Jung & Camarena, 2025), reducing their ability to map knowledge onto technology, recognise its usefulness (Wang et al., 2023), or integrate it effectively into workflows. In this way, high technostress disrupts the pathway from AI literacy to perceived fit. Conversely, low levels of technostress indicate that environmental demands are manageable. Under such conditions, employees can fully leverage their collaborative AI literacy to explore system functionalities, troubleshoot issues and adapt technology to their tasks. This unimpeded use of AI capabilities strengthens task–technology alignment and enhances performance. Therefore, we propose a conditional effect in which the positive relationship between collaborative AI literacy and TTF depends on the level of technostress; thus, we reached to the below given hypothesis and conceptual framework presented as Figure 1:
H3: Technostress moderates the relationship between collaborative AI literacy and task–technology fit, such that the relationship is weaker when technostress is high.
Methodology
Research design and context
In this research, a two-wave, multi-source survey design was used to test the hypothesised model of collaborative AI literacy with creative performance and technology-enabled performance mediated by TTF, in the presence of technostress as a critical moderator. The predictor and intermediate variables were measured at Time 1, and the outcome variable at Time 2, a design feature that helps to reduce common method bias and allows making more causal inferences concerning the hypothesised relationships (Podsakoff et al., 2003). China represents a strategic and ideal setting for this inquiry, as it is both the world’s largest market for industrial nations pursuing aggressive, state-led initiatives in AI and advanced manufacturing (Liu & Dunford, 2016). Data were collected from China’s leading services hubs – Dongguan, Shenzhen, Shanghai and Beijing – to ensure generalisability within the core of the nation’s advanced manufacturing sector (Liu & Dunford, 2016). Focusing on these cities allowed us to capture a wide spectrum of employee roles, from hardware integration to software intelligence, thereby enhancing the external validity of our findings.
Sample and data collection procedure
In this study, the target population was the full-time employees who are directly involved in the design, development and production. In order to reach this population, formal relationships were established with the Human Resources (HR) departments of various companies in the four cities chosen. The data-collection procedure has been carefully designed to maintain the rigour of the methodology. Measurement scales were all originally in English and were translated using the Brislin (1970) back-translation method into Mandarin Chinese to ensure semantic and conceptual correctness. The pilot test was then performed on 15 employees to ensure clarity and contextual appeal, and some slight wording changes were carried out. In Time 1 (T1), a survey through invitation emails and WeChat with a survey link was sent to 750 valid employees. The collaborative AI literacy, TTF, technology-enabled performance, technostress, and demographic controls were reflected in the T1 survey. To ensure attentiveness, an instructed response item was included as an attention check (Meade & Craig, 2012). To maintain confidentiality, employees used an anonymous code that was generated by the organisation to match. In 3 weeks with a reminder in between, 512 full responses were obtained with a response rate of 68.3%.
Three weeks after T1 closure, immediate direct report supervisors of the 512 respondents were requested to fill out the Time 2 (T2) assessment, giving creative performance ratings through the identical anonymous code. This time distance between predictors and effects was a procedural solution to the issue of common method variance (Podsakoff et al., 2003). From these, 410 matched supervisor ratings were retrieved, which on success of the attention check, 403 valid matched dyads were obtained, which was an overall effective response rate of 53.7 (see Table 1). They were not forced to participate in the survey, and informed consent was obtained online, and then the respondents were allowed to access the survey. In order to protect anonymity, no personally identifiable information was received, and all matching was done with anonymous employee codes. The secure storage of data was done within a password-protected server in compliance with the institutional data protection rules, so that the data remain confidential during the entire research process.
| TABLE 1: Demographic profile of the sample (N = 403). |
Measures
All constructs were measured using established multi-item scales adapted to the context of AI-enabled work. Responses were collected on five-point Likert scales. All scales demonstrated strong reliability and validity in the current sample, as reported in the results section. Collaborative AI literacy was measured using the 13-item scale developed by Sidra and Mason (2025). The scale assesses an individual’s competency to effectively evaluate, use, and understand the ethical implications of collaborative AI tools. A sample item includes thinking about your knowledge and experience of AI systems, how you rate your ability to: ‘determine what information the AI needs to perform the task appropriately’. Responses were collected on a 5-point Likert scale ranging from 1 (very poor) to 5 (excellent). Task-technology fit was measured using a 3-item scale developed by Goodhue and Thompson (1995). The scale was adapted from Larsen et al. (2009) and included the following items: ‘Using this technology fits with the way I work’, ‘Using this technology does not fit with my practice preferences’ (reverse-coded) and ‘Using this technology fits with my work practice’. Responses were captured on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).
Technology-enabled performance was measured using a 3-item scale adapted from Tarafdar et al. (2015), which has also been used in the study by Nuutinen et al. (2022). The original items were adapted to reflect performance. The items assessed the extent to which employees perceive technology as beneficial for enhancing their work performance and collaboration. A sample item is: ‘using technology helps me achieve better results in my tasks’. Responses were captured on a five-point Likert scale ranging from 1 (totally disagree) to 5 (totally agree). Technostress was measured using the techno-complexity dimension with 5-items developed by Tarafdar et al. (2007), following prior research demonstrating that single-dimension measurement is both theoretically acceptable and empirically supported (Borle et al., 2021; Fischer & Riedl, 2017; Kelderman, 2025; Ragu-Nathan et al., 2008). In technology-intensive contexts, techno-complexity is the most defensible moderator, as it captures employees’ perceived skill inadequacy and cognitive difficulty when managing complex AI systems, making it theoretically coherent for examining the relationship between AI adoption and creative self-efficacy. A sample item includes ‘I often find it too complex for me to understand and use new technology’. Respondent rated each time on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree).
Creative performance was measured using a 6-item scale adapted from Wang and Netemeyer (2004). The scale, validated in prior studies as well (e.g. Chiu et al., 2023), was used to assess employees’ ability to generate and implement innovative ideas. Direct supervisors rated their employees’ performance on a five-point Likert scale (1 = ‘strongly disagree’ to 5 = ‘strongly agree’). A sample item is: ‘this employee comes up with new ideas for solving technical problems’. Control variables consistent with prior research, age, gender, education, tenure and work experience were included as control variables because of their potential influence on performance outcomes (Tierney & Farmer, 2002).
Analytical strategy
The analysis proceeded in three sequential stages: (1) data preparation and validation, (2) measurement model evaluation and (3) structural model testing. Analyses were conducted using statistical package for the social sciences (SPSS) 25 and analysis of moment structures (AMOS) 24.
Data screening and common method bias
Prior to model testing, data were screened for missing values, outliers (Mahalanobis distance), normality (skewness and kurtosis within ±2) and multicollinearity (variance inflation factor [VIF] < 5). To address common method bias, we employed both procedural remedies (multi-source, two-wave design) and statistical controls. A Harman’s single-factor test indicated that no single factor explained the majority of the variance. Additionally, a common latent factor test in AMOS confirmed that common method variance did not significantly distort the measurement model.
Measurement model validation
Confirmatory factor analysis (CFA) was conducted to assess the psychometric properties of the scales. Convergent validity was supported by standardised factor loadings > 0.60, composite reliability (CR) > 0.70, and average variance extracted (AVE) > 0.50 for all constructs. Discriminant validity was established using the Fornell-Larcker criterion, where the square root of each construct’s AVE exceeded its correlations with other constructs. Model fit was evaluated using widely accepted thresholds: Comparative Fit Index (CFI)/Tucker–Lewis Index (TLI) > 0.90, Root Mean Square Error of Approximation (RMSEA) < 0.08, Root Mean Square Residual (SRMR) < 0.08. Alternative nested models were compared to confirm the superiority of the hypothesised five-factor structure.
Structural model and hypothesis testing
After validating the measurement model, structural equation modelling (SEM) was used to test the hypothesised relationships. Direct, indirect and moderated effects were estimated using maximum likelihood estimation. Mediation was tested using bias-corrected bootstrapping with 5000 resamples to generate 95% confidence intervals (CIs). Moderation was examined by creating a mean-centred interaction term (Collaborative AI Literacy × Technostress) and including it in the structural model. Simple slopes analysis was conducted to probe significant interactions at high (+1 standard deviation [SD]) and low (–1 SD) levels of technostress.
Moderated mediation test
To formally test the moderated mediation model, we calculated the Index of Moderated Mediation (IMM) using bootstrapping procedures recommended by Hayes (2015). The IMM quantifies how the indirect effect of collaborative AI literacy on performance outcomes (via TTF) changes across levels of technostress. Statistical significance of the IMM was determined using bootstrapped CIs (5000 samples). This approach provides a robust and transparent test of conditional indirect effects, directly addressing the editorial concern regarding moderated mediation reporting.
Ethical considerations
Ethical clearance to conduct this study was obtained from the National College of Business Administration and Economics and the Institutional Review Board. (Ref. No. 2025-A130).
Results
The initial screening of data was on values missing, abnormal, normality and multicollinearity cases. The few missing entries that were found have been corrected through the mean-substitution process in SPSS. The outlier detection was performed using the stem-and-leaf method, which detected four suspicious cases, which were then filtered out, and only 403 valid responses were retained. The rest of the data distribution met acceptable standards of normality with skewness and kurtosis equal to +1 and +3, respectively. In line with Ringle et al. (2015), the VIF values were used to measure multicollinearity, with all the values comfortably below the advisable 3-cutoff.
Factor loading
The analysis of the standard factor loadings showed significant and high relationships among all items observed and the latent constructs. Maximum likelihood estimation was used to analyse these loadings by CFA to evaluate the degree to which each of the measured variables represented its intended construct. Collaborative AI literacy has a loading range of between 0.675 and 0.755. In the case of Technostress, five of the items showed very large loadings (between 0.815 and 0.931); two preliminary items were eliminated before this analysis because of loadings less than the 0.50 criterion (Hair et al., 2010), a common method of purifying a scale that only retains the strongest items (Kline, 2023). In the same manner, two items were dropped out of the TTF construct, which then showed loadings of between 0.630 and 0.776 on the latter three items. Lastly, the items of both performances had a range between 0.624 and 0.824. All the retained values are well above the accepted minimum, which proves that each of the indicators is measuring what its theory should measure (Hair et al., 2010; Kline, 2023). Item removal decisions were guided by established measurement criteria (Hair et al., 2010; Kline, 2023). Although some items were dropped because of low factor loadings, the retained indicators preserved the conceptual domain of each construct and demonstrated strong reliability and validity.
Confirmatory factor analysis for validity, composite reliability and model fit
The measurement model showed a high level of convergent and discriminant validities (see Table 2). Convergent validity was also supported because all constructs had CR above the recommended 0.70 limit and AVE above 0.50 (Hair et al., 2010). Composite reliability and AVE values in this paper were 0.817–0.951, which means that there was good internal consistency and convergence among the items. The Fornell-Larcker criterion (1981) was used to validate the discriminant validity. The AVE (diagonal values) square root of each construct was higher than its correlations with the other constructs (off-diagonal values). For example, the square root of AVE for collaborative AI literacy (0.777) was higher than its correlations with creative performance (0.174), TTF (0.340), technology-enabled performance (0.124) and technostress (–0.037). This pattern was consistent across all latent variables, confirming that each construct is empirically distinct and captures a unique conceptual domain.
| TABLE 2: Composite reliability, convergent and discriminant validity. |
We follow Hair et al. (2010) to further establish construct distinctiveness. Table 3 presents the alternative CFA models that were estimated in which collaborative AI literacy items were combined with TTF and performance constructs. These models demonstrated significantly poorer fit compared to the hypothesised five-factor model, providing additional evidence of discriminant validity.
| TABLE 3: Fit statistics from measurement model comparison. |
Based on the descriptive statistics, it can be stated that the respondents reported rather high scores on collaborative AI literacy (M = 4.015, SD = 0.685), TTF (M = 4.258, SD = 0.507) and technology-enabled performance (M = 4.426, SD = 0.477). The creative performance has been medium (M = 3.828, SD = 0.773), and the technostress has been at a low level of mean (M = 2.631, SD = 1.477). The higher SD of technostress indicates that the perceptions of technostress experienced by employees had a general dispersion; on the other hand, the results had more agreements on the values of the other constructs, which can be viewed as having a high level of consensus. Internal consistency was high as indicated by Cronbach alpha coefficients of 0.816–0.951, which is way above the set 0.70 (Nunnally & Bernstein, 1994). These results indicate that the measurement scales are dependable and internally consistent.
The correlation analysis indicates a consistent and theoretically consistent pattern of correlations between the study variables. The calculation of partial correlations was carried out with the control of demographic variables (age, gender, education, experience). The positive correlations between collaborative AI literacy and TTF (r = 0.287, p < 0.001), technology-enabled performance (r = 0.136, p < 0.01), and creative performance (r = 0.174, p < 0.001) were all found to be significant and positive. Task-technology fit too showed significant positive correlations with technology-enabled performance (r = 0.165, p < 0.001) and creative performance (r = 0.158, p < 0.01), which further proves it as an enabler of technological efficiency and creative behaviour. Conversely, technostress showed weak, largely negative relationships with the other variables, some marginally significant, indicating the possibility that it may just be slightly detrimental to performance and creativity, but not on a strong bivariate level. The significance of these correlations is insignificant enough to suggest that the action of technostress can take more complicated forms, i.e. moderation in our case, as opposed to direct correlations. All in all, the descriptive statistics, reliability measures and correlation patterns taken as a whole represent a solid empirical base of the hypothesised relationships and the conceptual framework of the study model (see Table 4). The observed correlations between collaborative AI literacy and TTF (r = 0.287), and between TTF and creative performance (r = 0.158), are statistically significant but small-to-modest by conventional standards (Cohen, 2013). These effect sizes suggest that while TTF is a meaningful mediator, a substantial portion of variance in performance outcomes remains unexplained by the current model. Readers should avoid overinterpreting the practical magnitude of the indirect effects reported in Table 6.
| TABLE 4: Descriptive, reliability and correlation analyses (N = 403). |
Moderated mediation results
Table 5 presents the regression results indicating collaborative AI literacy positively predicts TTF (β = 0.219, p < 0.001), and this effect is significantly weakened by technostress (β = -0.056, p = 0.012), showing that technostress functions as a boundary condition. Task–technology fit, in turn, positively influences both technology-enabled performance (β = 0.124, p = 0.011) and creative performance (β = 0.169, p = 0.029). The direct effect of collaborative AI literacy on creative performance remains significant (β = 0.158, p = 0.006), indicating partial mediation, whereas its effect on technology-enabled performance is not significant (β = 0.057, p = 0.113), suggesting full mediation through TTF. These results collectively demonstrate the presence of first-stage moderated mediation. Specifically, the mediation of TTF in the relationship between collaborative AI literacy and performance outcomes is conditional on the level of technostress. At lower levels of technostress, AI literacy more effectively translates into TTF, which subsequently enhances both creative and technology-enabled performance. At higher levels of technostress, the positive effect of AI literacy on TTF is reduced, weakening the indirect effect on performance outcomes.
| TABLE 5: Regressions weights for structural paths. |
The conditional indirect effects and the IMM were calculated using the product of coefficients method, based on the path estimates from the structural model, following the procedures outlined by Hayes (2015). The moderated mediation analysis, presented in Table 6, provides formal statistical evidence for the hypothesised conditional indirect effects. Analysis reveals that technostress significantly reduces the positive indirect effect of collaborative AI literacy on performances through TTF. The IMM is negative and significant for both technology-enabled performance (IMM = -0.007) and creative performance (IMM = -0.009), confirming that higher technostress weakens these mediated pathways. While the indirect effects show a clear decreasing pattern from low (β = 0.034 and 0.047) to high (β = 0.020 and 0.027) technostress levels, the statistical significance at high technostress depends on the precision of the estimate, as indicated by the bootstrap CIs. The overall pattern demonstrates that the performance benefits of AI literacy are substantially diminished under conditions of high technostress.
| TABLE 6: Moderated mediation analysis: Conditional indirect effects at low, mean and high levels of technostress. |
Figure 2 shows that collaborative AI literacy improves TTF, but this effect changes sharply depending on the level of technostress. When technostress is low, the slope is positive and reasonably steep, indicating that employees can convert higher AI literacy into a stronger sense of TTF. When technostress is high, the slope becomes flat and slightly negative. This produces a crossover pattern: under high technostress, increases in AI literacy do not enhance TTF and may even relate to a small decline. Overall, the graph indicates that technostress not only weakens but can reverse the positive influence of AI literacy on TTF. This aligns with the significant negative interaction term (β = –0.056, p = 0.02) and suggests that in high-stress environments, greater AI literacy may heighten awareness of technology–task misfit. A practical implication is that reducing technostress is essential for employees to benefit from AI literacy. In high-stress conditions, enhancing AI skills alone may not improve perceived fit and could have unintended negative effects if stress is not addressed.
 |
FIGURE 2: Moderating effects of technostress on the relationship of collaborative artificial intelligence literacy and task technology fit. |
|
Discussion
This study explored how collaborative AI literacy affects employee performance in AI-enabled workplaces. Grounded in TTF theory (Goodhue & Thompson, 1995), we tested a moderated mediation model in which TTF mediates the relationship between collaborative AI literacy and two performance outcomes – technology-enabled performance and creative performance – with technostress as a boundary condition. A two-wave, multi-source survey provided strong support for the hypothesised model. The results clarify when the relationship between AI literacy and performance is strengthened or weakened. Consistent with TTF theory, higher collaborative AI literacy was associated with stronger TTF, which in turn positively predicted both performance outcomes. Notably, TTF fully mediated the relationship between collaborative AI literacy and technology-enabled performance, but only partially mediated the relationship with creative performance. This suggests that AI literacy drives routine efficiency through perceived fit, while also stimulating creativity through additional pathways.
Theoretical contributions
Firstly, it extends TTF theory by identifying collaborative AI literacy as a strategic antecedent of TTF. While traditional TTF research focuses on the alignment between task demands and technology features, our findings demonstrate that employees’ capability to collaborate effectively with AI systems directly shapes perceived fit. This reframes TTF as a dynamic, user-driven construct, highlighting that managers must consider employee competencies – not just technology design – when seeking performance gains. Secondly, the study clarifies distinct mediation pathways. The full mediation of technology-enabled performance indicates that routine efficiency critically depends on perceived fit, suggesting that managers should prioritise interventions that enhance TTF to optimise operational outcomes. In contrast, the partial mediation for creative performance shows that collaborative AI literacy can independently stimulate innovation, implying that skill development can generate value beyond immediate task alignment, enabling managers to foster both efficiency and creativity simultaneously.
Thirdly, the study integrates a boundary condition via technostress, showing that perceived fit and the benefits of AI literacy are conditional on cognitive and emotional states. High levels of techno-complexity can diminish the positive impact of literacy, indicating that managers must monitor stress and workload when implementing AI programmes. This underscores that investments in AI literacy alone may not yield performance gains under high-stress conditions, and contextualised strategies are necessary to ensure alignment between human capabilities and AI systems. Finally, the research advances the conceptualisation of socio-technical competence, emphasising collaborative AI literacy as essential for modern AI-enhanced workplaces. Unlike traditional digital or computer literacy, collaborative AI literacy captures employees’ ability to integrate AI recommendations into workflows, make context-sensitive decisions, and work effectively alongside intelligent systems. This highlights that managers should view workforce capability as an organisational asset, central to both AI adoption and innovation, rather than focusing solely on technology implementation, as suggested by Winkler-Titus et al. (2026). Finally, the moderated mediation analysis reveals that technostress significantly weakens the indirect effect of AI literacy on performance via TTF. This extends the job demands-resources (JD-R) and transactional stress theories by identifying technostress as a critical boundary condition that constrains the translation of technology-related resources into performance outcomes.
Practical implications for managers and decision-makers
The findings offer a clear, evidence-based roadmap for organisations seeking to leverage AI for performance without undermining employee well-being. Managers and HR leaders should move beyond isolated AI training and adopt an integrated capability–environment strategy. Specifically, three actionable decisions emerge: Redesign AI Training to Include Technostress Resilience. (1) Decision: Integrate stress-awareness and cognitive-coping modules into AI literacy programmes. Action: Instead of standalone technical workshops, develop blended learning paths that teach employees how to manage complexity, uncertainty, and cognitive overload when collaborating with AI. This ensures literacy translates into confidence rather than frustration. Adopt ‘Fit-Centric’ Implementation Protocols. (2) Decision: Make TTF a leading indicator in AI rollouts. Action: Prior to scaling AI tools, conduct pilot phases where TTF is systematically measured through employee feedback. Use this data to co-design workflows, simplify interfaces and customise system features – treating fit as a dynamic metric rather than a fixed assumption. Foster Low-Stress Digital Work Environments Through Leadership and Design. (3) Decision: Position technostress reduction as a managerial key performance indicator (KPI), not just an IT issue. Action: Equip line managers to recognise signs of techno-complexity and provide real-time support. Simultaneously, mandate that AI systems adhere to human-centric design principles – minimising steps, clarifying AI decision logic and providing ‘explainability’ features to reduce cognitive friction.
For senior leaders, the study underscores that AI performance is not an IT project but an organisational change initiative. Investments in AI tools and training will underdeliver unless paired with deliberate efforts to reduce digital stress and enhance tool–task alignment. By adopting the integrated framework validated here, organisations can transform AI potential into sustained gains in efficiency, innovation and employee adaptability.
Limitations and future directions
This research has limitations despite the contributions that can be made and must be taken into account when interpreting the results and designing future studies. Firstly, the sample was selected might not be generalisable to other industries, professional jobs and cultural conditions with varying degrees of AI maturity, organisational standards and perceptions of stress. To prove the validity of boundaries and strengthen the external validity, future work should challenge this model on both sector (e.g. healthcare, education) and cultural levels. Secondly, though the scales were validated, there are a number of measurement limitations. Although this study adopts a focused operationalisation of technostress and removes several scale items during validation, these decisions were theoretically grounded and consistent with best practices in measurement purification. Future research may extend this model by incorporating multiple technostress dimensions. Also, the creative performance was based entirely on the supervisor ratings, and this might not be enough to reflect the daily creative behaviours. Future research would gain advantages of multidimensional measurement – disaggregating AI literacy, a comprehensive technostress scale and a multi-source or objective performance measure.
Thirdly, a two-wave design is superior to cross-sectional data, but it does not provide a well-grounded causality. Reverse or reciprocal effects – in which the performance determines later fit or literacy – cannot be eliminated. Stronger longitudinal designs that incorporate three or more waves (or field experiments that manipulate AI literacy training and support) would enable more robust causal inferences and dynamic changes in these constructs with time. Lastly, the model only tested one mediator, as well as one moderator and did not test other theoretically important mechanisms. Notable variables like AI trust, team psychological safety, leadership support or personal characteristics, like learning agility, can further explain how literacy can be converted into performance. Future studies ought to explore parallel or consecutive mediation routes and other moderators to develop a more sophisticated perspective of the AI literacy-performance association. Specifically, the identified crossover effect between literacy and technostress is a phenomenon that should be qualitatively examined in order to identify the reasons why high-stress conditions lead knowledgeable users to feel a poorer fit, either because of an increase in critical awareness or cognitive depletion or underperformance.
Conclusion
This study demonstrates that collaborative AI literacy improves both technology-enabled and creative performance primarily through TTF, yet this positive pathway is conditional on technostress levels. Under high technostress, the benefits of AI competency can be significantly reduced or even reversed, highlighting technostress as a critical boundary condition and challenging the assumption that increased AI literacy automatically improves outcomes. By empirically validating a moderated mediation model grounded in TTF theory, the study provides a decision-oriented framework for understanding how AI integration impacts employee performance. The findings underscore that managers cannot rely solely on literacy initiatives; they must implement integrated strategies that enhance task–technology alignment, proactively monitor and mitigate stress, and design AI systems as supportive, cognitively manageable tools. All four hypotheses were supported and therefore accepted. H1, proposing a positive relationship between collaborative AI literacy and TTF, was accepted. H2a and H2b, proposing the mediating role of TTF for technology-enabled performance and creative performance, respectively, were both accepted. H3, proposing the moderating role of technostress, was also accepted. Overall, the study offers practical guidance for leveraging AI investments effectively, showing that sustainable performance arises from a balanced socio-technical approach – where skill development, system design and stress management converge to maximise both operational efficiency and innovative capacity in AI-driven workplaces. The modest effect sizes (e.g. conditional indirect effects ranging from 0.020 to 0.046) indicate that collaborative AI literacy and TTF explain only a small portion of the variance in performance outcomes. Practitioners should view AI literacy as one of several levers – alongside job design, leadership support and psychological safety – rather than a standalone solution for performance improvement.
Acknowledgements
The authors thank the employees and supervisors who participated in the two-wave survey, as well as the HR departments of the participating companies in Dongguan, Shenzhen, Shanghai, and Beijing for facilitating data collection. This research was supported by the Hunan Key Laboratory of Intelligent Decision-making Technology for Emergency Management, 2020TP1013.
Competing interests
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
CRediT authorship contribution
Wei Tan: Conceptualisation, Formal analysis, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing. Muhammad Nawaz: Conceptualisation, Supervision, Writing – review & editing. Tong Shu: Conceptualisation, Supervision, Writing – review & editing. Beenish Ramzan: Formal analysis, Methodology, Writing – review & editing. All authors reviewed the article, contributed to the discussion of results, approved the final version for submission and publication, and take responsibility for the integrity of its findings.
Funding information
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Data availability
The data that support the findings of this study are available from the corresponding author, Muhammad Nawaz, upon reasonable request.
Disclaimer
The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency or that of the publisher. The authors are responsible for this article’s results, findings and content.
References
Aguirre-Urreta, M.I., & Marakas, G.M. (2025). The role of computer self-efficacy in task performance: Insights from two empirical studies. International Journal of Human–Computer Interaction, 41(19), 12325–12348. https://doi.org/10.1080/10447318.2025.2456539
Azpíroz-Dorronsoro, C., Fernández-Muñiz, B., Montes-Peón, J.M., & Vázquez-Ordás, C.J. (2024). Technostress and work-family conflict in ICT-user employees during the COVID-19 pandemic: The role of social support and mindfulness. Behaviour & Information Technology, 43(8), 1531–1553. https://doi.org/10.1080/0144929X.2023.2220051
Borle, P., Reichel, K., Niebuhr, F., & Voelter-Mahlknecht, S. (2021). How are techno-stressors associated with mental health and work outcomes? A systematic review. International Journal of Environmental Research and Public Health, 18(16), 8673. https://doi.org/10.3390/ijerph18168673
Brislin, R.W. (1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 1(3), 185–216. https://doi.org/10.1177/135910457000100301
Chakraborty, D., Troise, C., & Bresciani, S. (2025). Exploring consumer intentions to continue: Integrating task technology fit and social technology fit in generative AI based shopping platforms. Technovation, 142, 103189. https://doi.org/10.1016/j.technovation.2025.103189
Cheng, Y.M. (2020). Understanding cloud ERP continuance intention and individual performance: A TTF-driven perspective. Benchmarking: An International Journal, 27(4), 1591–1614. https://doi.org/10.1108/BIJ-05-2019-0208
Chin, T., Li, Q., Mirone, F., & Papa, A. (2025). Conflicting impacts of shadow AI usage on knowledge leakage in metaverse-based business models: A Yin-Yang paradox framing. Technology in Society, 81, 102793. https://doi.org/10.1016/j.techsoc.2024.102793
Chiu, Y.W., Amin, M.W., Li, S.T., & Ali, M. (2023). Spiritual leadership influence on employee creative service performance: A moderated mediation analysis. BMC psychology, 11(1), 262. https://doi.org/10.1186/s40359-023-01294-0
Cohen, J. (2013). Statistical power analysis for the behavioral sciences (2nd ed.). Routledge.
Coyne, J.C., & Holroyd, K. (1982). Stress, coping, and illness. In T. Millon, C.J. Green, & R.B. Meagher (Eds.), Handbook of clinical health psychology (pp. 103–127). Springer. https://doi.org/10.1007/978-1-4613-3412-5_6
Fischer, T., & Riedl, R. (2017). Technostress research: A nurturing ground for measurement pluralism? Communications of the Association for Information Systems, 40(1), 375–401. https://doi.org/10.17705/1CAIS.04017
Fornell, C., & Larcker, D.F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
Goodhue, D.L., & Thompson, R.L. (1995). Task-technology fit and individual performance. MIS Quarterly, 19(2), 213–236. https://doi.org/10.2307/249689
Hair, J.F., Anderson, R.E., Babin, B.J., & Black, W.C. (2010). Multivariate data analysis: A global perspective (Vol. 7). Pearson.
Hayes, A.F. (2015). An index and test of linear moderated mediation. Multivariate Behavioral Research, 50(1), 1–22. https://doi.org/10.1080/00273171.2014.962683
Jankelova, N., Dabić, M., & Halaszovich, T. (2025). Organizational climate: Citizenship behavior, idiosyncratic deals and leader–member exchange. Management Decision, 63(13), 600–630. https://doi.org/10.1108/MD-01-2025-0012
Janssen, M., Van der Voort, H., & Wahyudi, A. (2017). Factors influencing big data decision-making quality. Journal of Business Research, 70, 338–345. https://doi.org/10.1016/j.jbusres.2016.08.007
Jia, N., Luo, X., Fang, Z., & Liao, C. (2024). When and how artificial intelligence augments employee creativity. Academy of Management Journal, 67(1), 5–32. https://doi.org/10.5465/amj.2022.0426
Jiang, D., Nawaz, M., Hanif, M.W., & Liu, S. (2026). Grey relational analysis and regression-based structural equation modeling of the extrarole performance of healthcare professionals. Social Behavior and Personality, 54(2), e14493. https://doi.org/10.2224/sbp.14493
Jung, H., & Camarena, L. (2025). Street-level bureaucrats & AI interactions in public organizations: An identity based framework. Public Performance & Management Review, 48(6), 1423–1452. https://doi.org/10.1080/15309576.2024.2447352
Kelderman, S. (2025). The relationship between technostress and job performance: The mediating role of work engagement. Tilburg University.
Khan, A.N., Soomro, M.A., & Pitafi, A.H. (2025a). AI in the workplace: Driving employee performance through enhanced knowledge sharing and work engagement. International Journal of Human–Computer Interaction, 41(17), 10699–10712. https://doi.org/10.1080/10447318.2024.2436611
Khan, S., Bhatti, G.A., Nawaz, M., Xie, N., & Zhang, W. (2025b). Forecasting ICT service export trends using advanced grey models. Journal of Grey System, 37(6), 66. Retrieved from https://jgrey.nuaa.edu.cn/EN/Y2025/V37/I6/66
Khan, S., Khan, M.J., Nawaz, M., Liu, S., & Ramzan, B. (2026). Modeling the moderation: A comparative application of regression and grey relational analysis in healthcare. Health Services and Outcomes Research Methodology, 1–19. https://doi.org/10.1007/s10742-026-00372-1
Kline, R.B. (2023). Principles and practice of structural equation modeling. Guilford Publications.
Klopping, I.M., & McKinney, E. (2004). Extending the technology acceptance model and the task-technology fit model to consumer e-commerce. Information Technology, Learning & Performance Journal, 22(1), 35–48.
Kumar, A., Krishnamoorthy, B., & Bhattacharyya, S.S. (2024). Machine learning and artificial intelligence-induced technostress in organizations: A study on automation-augmentation paradox with socio-technical systems as coping mechanisms. International Journal of Organizational Analysis, 32(4), 681–701. https://doi.org/10.1108/IJOA-01-2023-3581
Larsen, T.J., Sørebø, A.M., & Sørebø, Ø. (2009). The role of task-technology fit as users’ motivation to continue information system use. Computers in Human Behavior, 25(3), 778–784. https://doi.org/10.1016/j.chb.2009.02.006
Lazarus, R.S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer.
Lee, Y., Ahn, H., & Jo, H. (2025). Exploring the drivers and outcomes of generative AI usage: The role of metacognition. International Journal of Human–Computer Interaction, 42(4), 2098–2120. https://doi.org/10.1080/10447318.2025.2526592
Li, L., & Wang, X. (2021). Technostress inhibitors and creators and their impacts on university teachers’ work performance in higher education. Cognition, Technology & Work, 23(2), 315–330. https://doi.org/10.1007/s10111-020-00625-0
Liu, W., & Dunford, M. 2016. Inclusive globalization: Unpacking China’s belt and road initiative. Area Development and Policy, 1(3), 323–340. https://doi.org/10.1080/23792949.2016.1232598
Matsepe, N.T., & Van der Lingen, E. (2022). Determinants of emerging technologies adoption in the South African financial sector. South African Journal of Business Management, 53(1), 2493. https://doi.org/10.4102/sajbm.v53i1.2493
McKinsey & Company. (2023). The state of AI in 2023: Generative AI’s breakout year. McKinsey Digital. Retrieved from https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year
Meade, A.W., & Craig, S.B. (2012). Identifying careless responses in survey data. Psychological Methods, 17(3), 437–455. https://doi.org/10.1037/a0028085
Nawaz, M., & Liu, S. (2025). Do good have good! Idiosyncratic deals and creative performance through the lenz of cyberloafing and deontic justice: A moderated mediation model. International Journal of Healthcare Management, 19(2), 323–336. https://doi.org/10.1080/20479700.2025.2466536
Nawaz, M., Liu, S., Xie, N., & Ramzan, B. (2025). Evaluation of barriers to artificial intelligence adoption: Grey multi-criteria decision-making. Grey Systems: Theory and Application, 15(4), 732–754. https://doi.org/10.1108/GS-12-2024-0147
Nawaz, M., Liu, S., Xie, N., Atef, M., & Hanif, M.W. (2024). Evaluation of barriers to disabled elderly’s access to eHealth in China using grey relational analysis. Journal of Grey System, 36(5), 1–14. Retrieved from https://jgrey.nuaa.edu.cn/EN/Y2024/V36/I5/1
Nunnally, J.C., & Bernstein, I.H. (1994). Psychometric theory (3rd ed.). McGraw-Hill.
Nuutinen, S., Ahola, S., Eskelinen, J., & Kuula, M. (2022). How job resources influence employee productivity and technology-enabled performance in financial services: The job demands–resources model perspective. Journal of Organizational Effectiveness: People and Performance, 9(2), 233–252. https://doi.org/10.1108/JOEPP-01-2021-0014
Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., & Podsakoff, N.P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi.org/10.1037/0021-9010.88.5.879
Pranteddu, L., Porlezza, C., Kuai, J., & Komatsu, T. (2024). From the ‘Desk Set’ to ‘Doraemon’: A comparative analysis on the sociotechnical imaginaries of artificial intelligence in news work. Global Media and China, 10(6), 20594364241278961. https://doi.org/10.1177/20594364241278961
Przegalinska, A., Triantoro, T., Kovbasiuk, A., Ciechanowski, L., Freeman, R.B., & Sowa, K. (2025). Collaborative AI in the workplace: Enhancing organizational performance through resource-based and task-technology fit perspectives. International Journal of Information Management, 81, 102853. https://doi.org/10.1016/j.ijinfomgt.2024.102853
Ragu-Nathan, T.S., Tarafdar, M., Ragu-Nathan, B.S., & Tu, Q. (2008). The consequences of technostress for end users in organizations: Conceptual development and validation. Information Systems Research, 19(4), 417–433. https://doi.org/10.1287/isre.1070.0165
Ringle, C.M., Wende, S., & Becker, J.M. (2015). SmartPLS 3. SmartPLS GmbH, Boenningstedt. Journal of Service Science and Management, 10(3), 32–49. Retrieved from http://www.smartpls.com
Sidra, S., & Mason, C. (2025). Generative AI in human–AI collaboration: Validation of the collaborative AI literacy and collaborative AI metacognition scales for effective use. International Journal of Human–Computer Interaction, 42(7), 5084–5108. https://doi.org/10.1080/10447318.2025.2543997
Tarafdar, M., Pullins, E.B., & Ragu-Nathan, T.S. (2015). Technostress: Negative effect on performance and possible mitigations. Information Systems Journal, 25(2), 103–132. https://doi.org/10.1111/isj.12042
Tarafdar, M., Tu, Q., Ragu-Nathan, B.S., & Ragu-Nathan, T.S. (2007). The impact of technostress on role stress and productivity. Journal of Management Information Systems, 24(1), 301–328. https://doi.org/10.2753/MIS0742-1222240109
Tierney, P., & Farmer, S.M. (2002). Creative self-efficacy: Its potential antecedents and relationship to creative performance. Academy of Management Journal, 45(6), 1137–1148. https://doi.org/10.5465/3069429
Wang, G., & Netemeyer, R.G. (2004). Salesperson creative performance: Conceptualization, measurement, and nomological validity. Journal of Business Research, 57(8), 805–812. https://doi.org/10.1016/S0148-2963(02)00483-6
Wang, H., Ding, H., & Kong, X. (2023). Understanding technostress and employee well-being in digital work: The roles of work exhaustion and workplace knowledge diversity. International Journal of Manpower, 44(2), 334–353. https://doi.org/10.1108/IJM-08-2021-0480
Wang, X., Xie, F., Cheng, G., & Shehzadi, K. (2025). Identifying and assessing the risks of artificial intelligence applications in smart libraries: Perspective of technostress. Information Development, 02666669251329032. https://doi.org/10.1177/02666669251329032
Winkler-Titus, N., Feirreira-Marcelino, N., & Kidd, M. (2026). Technology adoption and worker engagement in remote working: Insights on inclusion. South African Journal of Business Management, 57(1), 5455. https://doi.org/10.4102/SAJBM.v57i1.5455
Yang, B., & Sun, Y. (2025). Improving learning performance through AI-enabled task completing and network building: Integrating task-technology fit and social-technology fit perspectives. Interactive Learning Environments, 1–21. https://doi.org/10.1080/10494820.2025.2569777
Zhang, L., Shao, Z., Zhao, T., & Chen, K. (2025a). The influences of four dimensions of perceived fit on individuals’ utilisation of SPOCs: An extension of the task-technology fit model. Behaviour & Information Technology, 44(3), 611–629. https://doi.org/10.1080/0144929X.2024.2332449
Zhang, X., Qi, Z., Ma, L., & Zhang, G. (2025b). Assessing the curvilinear relationship in employee digital performance: A task-technology fit perspective. International Journal of Human–Computer Interaction, 41(4), 2615–2633. https://doi.org/10.1080/10447318.2024.2327181
|
Hide
Show all
Email this article (Login required)
