Psychological adaptation to generative AI in service work: knowledge elaboration and future orientation.
BMC psychology · 2026
Abstract originale
BACKGROUND: Generative AI (GenAI) is increasingly reshaping service work and creating new demands for employees' cognitive and psychological adaptation to human-AI interaction and collaboration. Although prior research has examined the organizational consequences of GenAI, limited attention has been paid to how employees' acceptance of GenAI is associated with service performance through cognitive adaptation and how this association varies across employees. Drawing on sociotechnical systems theory, this study examined whether knowledge elaboration serves as a cognitive adaptation mechanism linking GenAI acceptance to employee service performance, and whether future orientation strengthens this process. METHODS: This study employed a two-wave, multi-source survey design. Data were collected from 890 supervisor-employee dyads in the service sector. Employees reported their GenAI acceptance, future orientation, and demographic information at Time 1, while employees reported knowledge elaboration and supervisors evaluated service performance at Time 2. The hypothesized mediation and moderated mediation models were tested using regression analysis and bootstrapping procedures. RESULTS: GenAI acceptance was positively associated with knowledge elaboration, which in turn was positively related to employee service performance. The estimated indirect association between GenAI acceptance and service performance through knowledge elaboration was significant, while the direct association remained significant. In addition, future orientation moderated the relationship between GenAI acceptance and knowledge elaboration, such that this relationship was stronger among employees with higher future orientation. The significant index of moderated mediation indicated that the estimated indirect association between GenAI acceptance and service performance through knowledge elaboration was stronger at higher levels of future orientation. CONCLUSION: The findings suggest that the association between GenAI acceptance and service performance involves employees' cognitive engagement with AI-generated knowledge and is stronger among employees with higher future orientation. This study advances understanding of psychological adaptation in AI-integrated work settings by highlighting knowledge elaboration as a key cognitive mechanism and future orientation as an important psychological boundary condition.
Parole chiave
Leggi il testo integrale originale
Testo estratto dalla pubblicazione: sono esclusi immagini, media, supplementi e parti soggette a diritti distinti.
Psychological adaptation to generative AI in service work: knowledge elaboration and future orientation
Background
Generative AI (GenAI) is increasingly reshaping service work and creating new demands for employees’ cognitive and psychological adaptation to human–AI interaction and collaboration. Although prior research has examined the organizational consequences of GenAI, limited attention has been paid to how employees’ acceptance of GenAI is associated with service performance through cognitive adaptation and how this association varies across employees. Drawing on sociotechnical systems theory, this study examined whether knowledge elaboration serves as a cognitive adaptation mechanism linking GenAI acceptance to employee service performance, and whether future orientation strengthens this process.
Methods
This study employed a two-wave, multi-source survey design. Data were collected from 890 supervisor–employee dyads in the service sector. Employees reported their GenAI acceptance, future orientation, and demographic information at Time 1, while employees reported knowledge elaboration and supervisors evaluated service performance at Time 2. The hypothesized mediation and moderated mediation models were tested using regression analysis and bootstrapping procedures.
Results
GenAI acceptance was positively associated with knowledge elaboration, which in turn was positively related to employee service performance. The estimated indirect association between GenAI acceptance and service performance through knowledge elaboration was significant, while the direct association remained significant. In addition, future orientation moderated the relationship between GenAI acceptance and knowledge elaboration, such that this relationship was stronger among employees with higher future orientation. The significant index of moderated mediation indicated that the estimated indirect association between GenAI acceptance and service performance through knowledge elaboration was stronger at higher levels of future orientation.
Conclusion
The findings suggest that the association between GenAI acceptance and service performance involves employees’ cognitive engagement with AI-generated knowledge and is stronger among employees with higher future orientation. This study advances understanding of psychological adaptation in AI-integrated work settings by highlighting knowledge elaboration as a key cognitive mechanism and future orientation as an important psychological boundary condition.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s40359-026-05477-3.
Introduction
Generative AI (GenAI) broadly refers to a class of AI systems capable of producing novel content—such as text, images, videos, and other media—in response to user prompts [1, 2]. Owing to its user-friendly interfaces, GenAI enables a wider range of individuals to engage in AI-supported tasks and problem solving with relative ease [3, 4]. As a result, GenAI is increasingly recognized as a practical means of improving productivity and enhancing work outcomes across diverse organizational contexts [5]. At the same time, the diffusion of GenAI introduces new cognitive and psychological adaptation demands for employees working in AI-enabled environments.
These adaptation demands are particularly salient in service firms, where value creation depends heavily on employees’ real-time problem solving, adaptive use of knowledge, and interactions with customers [3, 5]. In such settings, rapid and unpredictable technological change can be especially consequential. Employees often experience these changes as shifting job content, escalating skill requirements, and new human–AI collaboration demands, which challenge the sufficiency of existing knowledge, increase ambiguity, and require sustained cognitive and psychological adaptation [6]. By automating and augmenting tasks traditionally reliant on human judgment and creativity, GenAI is reshaping knowledge-intensive work and altering how knowledge is created, shared, and applied within organizations [1, 6]. These transformations are likely to have important implications for employee outcomes, particularly service performance. Given that employee service performance is vital to sustaining competitive advantage and long-term success in service industries [5], understanding how employees’ acceptance of GenAI relates to service performance is of clear theoretical and practical importance.
Despite the growing relevance of GenAI in organizational life, workplace research on employees’ adaptive responses to GenAI remains limited, particularly regarding whether knowledge elaboration helps explain the association between psychological acceptance of GenAI and service performance. Prior studies have linked GenAI to outcomes such as employee creativity [2], work productivity and quality [7], service innovation [8], perceived value [4], and employee expediency [9]. However, the psychological process through which employees’ acceptance of GenAI is associated with service performance remains insufficiently understood. This omission is important because, in service settings, employees’ ability to respond effectively to customer needs directly shapes organizational success.
A further gap concerns the mechanisms through which GenAI-related psychological responses are linked to employee outcomes. GenAI does not merely provide information; through interactive and conversational engagement, it can support employee learning by responding to user queries, offering coaching, providing contextual examples, and delivering in-depth explanations [1, 10]. However, the availability of these capabilities alone does not explain variation in employee service performance. Employees’ evaluations of GenAI as useful, easy to use, and manageable may shape the extent to which they cognitively engage with AI-generated knowledge. In this regard, knowledge elaboration may represent a cognitive adaptation mechanism through which employees make sense of, integrate, and apply AI-generated knowledge in service work. Knowledge elaboration refers to the extent to which individuals actively reflect on, interpret, integrate, and apply newly acquired knowledge in their work [11, 12]. It reflects a deeper cognitive process through which task-relevant information is exchanged, discussed, and transformed into actionable knowledge [12, 13]. Yet existing research rarely examines whether GenAI acceptance is linked to employee service performance through such knowledge-processing activities.
In addition, prior research has paid insufficient attention to the conditions under which the associations involving GenAI acceptance vary across employees. Specifically, the association between GenAI acceptance and knowledge elaboration—and the corresponding indirect association with service performance—may depend on individual psychological characteristics. One such characteristic is future orientation, defined as the extent to which individuals habitually focus on the future, think about long-term goals, and consider the future consequences of their current actions [14, 15]. As such, future orientation can be viewed as an adaptive psychological capacity that helps employees interpret technological disruption in terms of future learning and development opportunities. Accordingly, future orientation may shape the extent to which GenAI acceptance fosters knowledge elaboration and, in turn, service performance.
To address this issue, we examine knowledge elaboration as a mediating mechanism linking GenAI acceptance to employee service performance and future orientation as a boundary condition shaping this relationship. Accordingly, we ask: How is GenAI acceptance associated with employees’ service performance through knowledge elaboration as a cognitive adaptation process, and how does this relationship vary as a function of future orientation?
Taken together, these gaps point to three important yet insufficiently understood questions concerning employees’ adaptation to GenAI-enabled service work. Accordingly, this study addresses the following research questions:
RQ1: How is GenAI acceptance associated with employee service performance?
RQ2: Does knowledge elaboration function as a cognitive adaptation mechanism linking GenAI acceptance to employee service performance?
RQ3: How does future orientation condition the relationship between GenAI acceptance and knowledge elaboration and, consequently, the indirect relationship between GenAI acceptance and employee service performance?
To address these questions, we draw primarily on sociotechnical systems (STS) theory and complement it with the knowledge-based view to develop a moderated mediation model. STS theory proposes that work outcomes arise from the joint functioning and alignment of technical and social subsystems rather than from technology alone [16–18]. Accordingly, we conceptualize employees’ adaptation to GenAI-enabled service work as a sociotechnical and psychological process rather than a simple technological input–performance output relationship. In this model, GenAI represents the technical resource, while GenAI acceptance reflects employees’ psychological alignment with this resource in terms of its perceived usefulness, ease of use, and manageability. Knowledge elaboration captures the cognitive adaptation process through which employees interpret, integrate, and apply AI-generated knowledge in their service work, whereas future orientation represents an individual psychological capacity that strengthens this adaptation process. Employee service performance, in turn, constitutes a behavioral outcome of the effective alignment among technological resources, psychological acceptance, cognitive engagement, and individual adaptive capacity. Complementing this STS logic, the knowledge-based view suggests that technological resources generate value only when employees can interpret, integrate, and transform available knowledge into effective work actions [19]. Together, these perspectives enable us to explain how GenAI acceptance is associated with employee service performance through knowledge elaboration and why this process is stronger among employees with higher future orientation.
We test this model using a two-wave, multi-source survey of 890 supervisor–employee dyads in the service sector. By temporally separating the measurement of predictor and mediator variables and collecting service performance ratings from supervisors, this design helps reduce common method concerns and strengthens the basis for examining the proposed relationships [20]. This study makes three main contributions. First, it shows that employees’ adaptive responses to GenAI in service contexts are best understood as a sociotechnical and psychological process rather than as a direct technological effect [18]. Second, it extends the knowledge-based view by identifying knowledge elaboration as a cognitive adaptation mechanism through which GenAI acceptance is associated with employee service performance [1, 10, 19]. Third, it identifies future orientation as an adaptive psychological capacity that conditions the strength of the indirect association between GenAI acceptance and service performance through knowledge elaboration [14, 15]. These contributions also provide practical insight into how service organizations can derive greater value from GenAI by aligning technological implementation with employees’ acceptance, learning, knowledge use, and customer-facing work demands [21, 22].
Theory and hypotheses
Sociotechnical systems theory and the proposed model
STS theory posits that organizational outcomes arise from the joint functioning of technical and social subsystems rather than from either subsystem in isolation [16–18]. The technical subsystem comprises technologies, processes, and related infrastructure, whereas the social subsystem encompasses employees, teams, relationships, and the psychological and behavioral processes through which work is performed. Accordingly, the value of technology depends not only on its technical functionality, but also on how effectively it is integrated with employees’ cognitive and behavioral responses.
This perspective is particularly relevant to GenAI-enabled service work. GenAI constitutes a technical resource that provides suggestions, explanations, contextual examples, and problem-solving support [1, 10]. However, the availability of these resources alone does not necessarily correspond to higher service performance. In our model, GenAI acceptance reflects employees’ psychological alignment with the technical resource, whereas knowledge elaboration represents the cognitive process through which AI-generated knowledge is interpreted, evaluated, integrated, and incorporated into service actions. Employee service performance is not treated merely as an organizational performance indicator. Rather, it is conceptualized as the behavioral manifestation of employees’ psychological acceptance of GenAI and cognitive engagement with AI-generated knowledge in customer-facing work. This interpretation is consistent with research linking employee–AI interaction to frontline employees’ service behaviors and performance [23]. Future orientation, in turn, functions as an individual adaptive capacity that conditions the strength of the association between psychological acceptance and cognitive engagement.
This conceptualization is also consistent with research showing that workplace AI outcomes reflect the interplay among technological capabilities, employee responses, and implementation contexts [18, 24]. Complementing this STS perspective, the knowledge-based view explains why technological resources become relevant to work outcomes only when employees interpret, integrate, and apply the knowledge they provide [19].
GenAI acceptance and knowledge elaboration
GenAI acceptance is likely to be positively related to knowledge elaboration because employees who perceive GenAI as useful, easy to use, and manageable should be more willing to engage cognitively with its outputs. GenAI can respond to user queries, provide contextual examples and explanations, and generate alternative solutions in real time [1, 10]. However, employees are unlikely to elaborate on AI-generated knowledge unless they regard GenAI as a helpful and usable resource and invest cognitive effort in processing its outputs.
Knowledge elaboration refers to the extent to which individuals actively reflect on, interpret, integrate, and apply newly acquired knowledge in their work [11, 12]. In the present context, it represents a cognitive adaptation process through which employees clarify AI-generated information, evaluate its relevance, connect it with existing knowledge, and incorporate it into service practices.
At the construct level, GenAI acceptance reflects three closely related evaluative judgments. Perceived usefulness signals that AI-generated knowledge is relevant to employees’ work goals and has instrumental value, thereby motivating them to examine and apply it [25]. Perceived ease of use reduces the effort required to interact with GenAI, making sustained engagement with its outputs more feasible [25, 26]. Perceived manageability enhances employees’ sense of control over GenAI use and reduces uncertainty, a logic consistent with research on perceived external control in technology use [26]. Together, these judgments should increase employees’ willingness to engage deeply with AI-generated knowledge. Such engagement involves reflecting on, evaluating, verifying, refining, and integrating new information into existing knowledge structures [11–13].
From an STS perspective, GenAI acceptance therefore reflects employees’ psychological alignment with the technical resource, whereas knowledge elaboration represents the cognitive process through which AI-generated knowledge becomes integrated into service work [18, 27]. Accordingly, we propose:
Hypothesis 1
(H1). GenAI acceptance is positively related to knowledge elaboration.
Knowledge elaboration and employee service performance
Employee service performance depends heavily on employees’ ability to understand customer needs, respond appropriately in real time, and apply relevant knowledge effectively in customer interactions [21, 22]. In GenAI-enabled service settings, the mere availability of information is insufficient to ensure high performance because employees must respond to customer needs through context-sensitive judgment and action. Rather, they must be able to make sense of the information they receive and convert it into effective service actions, as prior research suggests that the value of digital information for customer service performance depends on its translation into localized capabilities and responsive service behaviors [11–13, 28]. Knowledge elaboration is particularly important in this regard because it reflects a cognitive adaptation process through which employees move beyond passive receipt of information and translate knowledge into contextually appropriate service actions.
Employees who engage in greater knowledge elaboration are more likely to interpret service-related information accurately, integrate new insights with existing expertise, and apply them flexibly to customer needs. Prior research suggests that deeper information and knowledge processing is associated with more effective integration of available knowledge into task performance and decision making [11, 13]. In addition, knowledge elaboration has been associated with more effective use of received knowledge and more favorable work-related outcomes [12]. Employees who engage in greater knowledge elaboration should therefore be better equipped to tailor their responses, address service problems, and provide context-sensitive service. Within the STS framework, knowledge elaboration represents the human-side integration process through which AI-generated outputs are incorporated into customer-facing service actions. Employee service performance, in turn, represents the behavioral outcome associated with this sociotechnical integration process [23]. Therefore, we propose:
Hypothesis 2
(H2). Knowledge elaboration is positively related to employee service performance.
The mediating role of knowledge elaboration
Taken together, the preceding arguments suggest that knowledge elaboration serves as a cognitive adaptation pathway linking GenAI acceptance to employee service performance. STS theory holds that technological resources generate value only when they are effectively integrated with human cognition and work practices [16–18]. In the present context, GenAI provides employees with informational resources, but these resources are unlikely to enhance service performance automatically. Employees who accept GenAI as useful, easy to use, and manageable should be more willing to engage with its outputs and integrate AI-generated knowledge into their work.
Knowledge elaboration involves reflecting on AI-generated information, evaluating its relevance, combining it with existing knowledge, and incorporating it into context-sensitive service actions. This reasoning is consistent with the knowledge-based view, which emphasizes that the value of available knowledge depends on its effective interpretation, integration, and application [19]. It is also consistent with evidence that digital information contributes to customer service performance when it is converted into localized capabilities and responsive service processes [28]. Thus, GenAI acceptance reflects employees’ psychological alignment with the technical resource, knowledge elaboration facilitates its cognitive integration into service practices, and service performance represents the behavioral outcome associated with this sociotechnical process. Accordingly, we propose:
Hypothesis 3
(H3). Knowledge elaboration mediates the relationship between GenAI acceptance and employee service performance.
The moderating role of future orientation
Although GenAI acceptance is expected to be positively related to knowledge elaboration, this relationship may vary across employees. Future orientation refers to the extent to which individuals habitually focus on long-term goals and consider the future consequences of their current actions [14, 15]. Future-oriented individuals tend to invest greater effort in learning and proactive behaviors whose benefits may emerge over time [29]. In GenAI-enabled service work, future orientation can therefore be understood as an individual adaptive capacity that shapes how employees respond to an evolving technical resource. Importantly, future orientation is expected not merely to promote knowledge elaboration directly, but to condition the extent to which employees’ psychological acceptance of GenAI develops into cognitive engagement with its outputs.
Specifically, employees with higher future orientation are more likely to view GenAI as a resource for long-term learning, skill development, and career adaptation rather than merely as a tool for immediate task completion. Consequently, when they accept GenAI as useful, easy to use, and manageable, they should be more willing to invest cognitive effort in interpreting, evaluating, and integrating its outputs. Recent qualitative research similarly indicates that employees proactively adapt their work practices in response to GenAI while considering its implications for skill development and career prospects [30]. By contrast, employees with lower future orientation may interpret GenAI acceptance primarily in terms of immediate task utility and invest less cognitive effort in elaborating on its knowledge outputs. Thus, among employees with comparable levels of GenAI acceptance, those with higher future orientation should engage more deeply with AI-generated knowledge than those with lower future orientation. From an STS perspective, future orientation thus strengthens the extent to which employees’ psychological alignment with the technical resource develops into cognitive integration. Therefore, we propose:
Hypothesis 4
(H4). Future orientation moderates the positive relationship between GenAI acceptance and knowledge elaboration, such that the relationship is stronger for employees with higher future orientation.
Extending this STS logic, future orientation should strengthen the indirect relationship between GenAI acceptance and employee service performance by increasing the extent to which employees convert psychological acceptance of the technical resource into knowledge elaboration. Thus, the proposed sociotechnical alignment process should be stronger among employees with higher future orientation. Accordingly, we propose:
Hypothesis 5
(H5). Future orientation positively moderates the indirect effect of GenAI acceptance on employee service performance through knowledge elaboration, such that the indirect effect is stronger for employees with higher future orientation.
Overall, Fig. 1 presents the proposed research model.
Method
Participants and procedure
Because this study focuses on workplace GenAI use, we collected data from employees in Chinese service-sector companies that had implemented GenAI tools. All sampled employees were using GenAI in their work at the time of data collection. We recruited supervisor–employee dyads from a variety of service-sector firms through professional and personal networks. Before data collection, we contacted supervisors in the participating firms and explained the purpose and procedures of the study. These supervisors were then asked to invite their immediate subordinates to participate.
To reduce potential common method bias, we adopted a two-wave, multi-source design and several procedural remedies [20, 31]. First, employees reported GenAI acceptance, future orientation, and demographic information at Time 1 and knowledge elaboration three months later at Time 2, providing temporal separation between the predictor and mediator measures. Second, service performance was rated by supervisors rather than by employees. Third, participants were assured that their responses would remain confidential, would be used only for research purposes, and would not be disclosed to their supervisors or organizations. Unique identification codes were used solely to match employee responses across waves with supervisor ratings.
Of the initial respondents, 54 did not participate at Time 2, 36 completed the Time 2 employee survey but could not be matched with a corresponding supervisor rating, and 30 were excluded because their Time 2 employee or supervisor responses were incomplete. These exclusion categories were mutually exclusive. The final sample consisted of 890 supervisor–employee dyads, representing a retention rate of 88.1%.
Following recommendations for assessing attrition in longitudinal research [32], we compared retained and excluded participants using the Time 1 variables available for both groups. Independent-samples t tests and chi-square tests revealed no significant differences in gender, age, education, tenure, GenAI acceptance, or future orientation (all ps > 0.05). These results suggest that systematic attrition was unlikely to have substantially biased the final sample. Table 1 presents the demographic characteristics of the sample.
Table 1Demographic information
Feature Category % N Feature Category % N
Gender Male 53.9 480 Age 20–30 years 33.7 300
Female 46.1 410 30–40 years 51.7 460
Tenure 1 year or below 7.5 67 40–50 years 12.4 110
1–2 years 15.4 137 50–60 years 2.2 20
3–4 years 21.3 190 Education Junior high school and below 0.7 6
5–7 years 17.3 154 High school 4.3 38
More than 7 years 38.4 342 Associate degree 13.5 120
Firm size Below 100 employees 21.1 188 Bachelor’s degree 54.0 481
100–300 employees 22.5 200 Graduate degree 27.5 245
300–500 employees 17.9 159 Industry Hotel and catering 44.9 400
500–1,000 employees 7.2 64 Tourism services 3.6 32
More than 1,000 employees 31.3 279 Retail 7.8 69
Firm age 5 years or below 10.8 96 Information technology 4.2 37
6–10 years 16.1 143 Financial services 7.5 67
11–15 years 15.4 137 Logistics and transportation 7.5 67
16–20 years 13.6 121 Education and training 5.8 52
More than 20 years 44.2 393 Others 18.7 166
Measures
All focal constructs were measured using established scales and assessed on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). We followed Brislin’s [33] translation and back-translation procedure to develop the Chinese questionnaire. GenAI acceptance and future orientation were reported by employees at Time 1, whereas knowledge elaboration was reported by employees and service performance was rated by supervisors at Time 2.
The GenAI acceptance measure was adapted from Davis [25] and contextualized for GenAI use based on Singh et al. [34] and Khan et al. [35]. Although all respondents had workplace experience with GenAI, the modal wording “would” was retained to preserve the evaluative framing of the adapted items. Accordingly, the measure captures experience-based evaluations of GenAI’s usefulness, ease of use, and manageability rather than the frequency or intensity of actual use. The knowledge elaboration items were adapted from Ling et al. [12] by replacing the original human knowledge source with GenAI. Future orientation was measured using Shipp et al.‘s [14] scale, whereas service performance was assessed using items from Liao and Chuang [21]. The complete measurement items, sources, standardized factor loadings, and reliability and validity statistics are presented in Supplementary Table S1.
At Time 1, employees also reported the control variables used in the analyses. Drawing on prior service-performance research [21, 22], we controlled for employee education and tenure, as well as firm size and firm age.
Analysis strategy
Descriptive statistics, scale reliability, and correlation analyses were conducted using SPSS 27.0, while the heterotrait–monotrait ratio (HTMT) was calculated using Mplus 8.0. H1, H2, and H4 were tested using OLS regression analyses, which are appropriate for examining relationships among observed composite variables [36]. H3 and H5 were tested using Hayes’ [37] PROCESS macro for SPSS, with Model 4 for mediation and Model 7 for moderated mediation. To statistically assess potential common method bias, we compared the hypothesized three-factor model for the employee-reported constructs with a single-factor model in which all GenAI acceptance, knowledge elaboration, and future orientation items loaded on one factor. Percentile bootstrap confidence intervals based on 5,000 resamples were used to assess the indirect effects and the index of moderated mediation. All reported regression coefficients are unstandardized coefficients (b). GenAI acceptance and future orientation were mean-centered before constructing the interaction term. Conditional effects were probed at the 16th, 50th, and 84th percentiles of future orientation, corresponding to FO values of 5, 6, and 7, respectively. These values represent relatively low, median, and relatively high observed levels rather than the conventional mean ± 1 SD values [37].
Results
Descriptive statistics
Table 2 presents the means, standard deviations, and correlations among the study variables. The zero-order correlations provide preliminary support for Hypotheses 1 and 2. Specifically, GenAI acceptance was positively correlated with knowledge elaboration (r = 0.483, p < 0.001), and knowledge elaboration was positively correlated with service performance (r = 0.467, p < 0.001).
Table 2Means, standard deviations and correlations for the study variables
Variables Mean SD 1 2 3 4 5 6 7
1.Education 4.035 0.801
2.Tenure 3.637 1.327 -0.002
3.Firm age 3.643 1.443 -0.032 0.188***
4.Firm size 3.052 1.548 0.058 0.184*** 0.507***
5.Service performance 5.769 1.088 -0.028 0.079* 0.021 0.053
6.GenAI acceptance 5.648 1.250 0.008 0.045 0.028 0.057 0.483***
7.Knowledge elaboration 5.069 1.308 0.019 -0.045 0.009 -0.001 0.467*** 0.483***
8.Future orientation 5.807 1.068 0.039 0.029 -0.001 0.034 0.556*** 0.554*** 0.457***
N = 890; *p < 0.05, **p < 0.01, ***p < 0.001
Measurement model
Before hypothesis testing, we conducted confirmatory factor analyses using the diagonally weighted least squares estimator. As shown in Table 3, the hypothesized four-factor model provided the best fit, χ²(84) = 664.540, p < 0.001, CFI = 0.988, TLI = 0.985, RMSEA = 0.081, and SRMR = 0.027. The four-factor model outperformed all alternatives, supporting construct distinctiveness. As shown in Supplementary Table S1, the standardized factor loadings ranged from 0.856 to 0.972 and were all significant. Cronbach’s alpha values ranged from 0.926 to 0.932, composite reliability values ranged from 0.953 to 0.956, and AVE values ranged from 0.805 to 0.877. Thus, all factor loadings and reliability estimates exceeded 0.70, and all AVE values exceeded 0.50, supporting satisfactory internal consistency and convergent validity.
Table 3Results of confirmatory factor analyses
Models χ² df χ²/df CFI TLI RMSEA SRMR
Four-factor model: GAI; KE; FO; SP 664.540 84 7.911 0.988 0.985 0.081 0.027
Three-factor model: GAI; KE; FO + SP 2366.318 87 27.199 0.952 0.942 0.172 0.088
Two-factor model: GAI; KE + FO+SP 4436.873 89 49.853 0.908 0.892 0.234 0.146
Single-factor model: GAI + KE+FO + SP 5931.678 90 65.907 0.877 0.856 0.270 0.192
GAI = GenAI acceptance; KE = knowledge elaboration; FO = future orientation; SP = service performance; CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. All χ² values were significant at p < 0.001. N = 890
As additional evidence of discriminant validity, we assessed the four focal constructs using the heterotrait–monotrait ratio (HTMT), which compares correlations across constructs with correlations among indicators of the same construct [38]. As shown in Table 4, the HTMT values ranged from 0.492 to 0.600 and were all below the recommended threshold of 0.85. These results support the empirical distinctiveness of GenAI acceptance, knowledge elaboration, future orientation, and service performance [36].
A three-factor model for the employee-reported constructs, χ²(51) = 547.602, CFI = 0.987, TLI = 0.983, RMSEA = 0.105, and SRMR = 0.028, substantially outperformed a single-factor model, χ²(54) = 3857.120, CFI = 0.900, TLI = 0.878, RMSEA = 0.281, and SRMR = 0.191. Thus, a single common factor was unlikely to account for the covariance among the employee-reported measures. Taken together with the procedural remedies, these results suggest that common method variance was unlikely to seriously affect the main findings.
Table 4Heterotrait–monotrait ratios among the focal constructs
Construct 1 2 3 4
1. GenAI acceptance —
2. Knowledge elaboration 0.519 —
3. Future orientation 0.597 0.492 —
4. Service performance 0.519 0.503 0.600 —
Values below the diagonal are heterotrait–monotrait ratios. All values are below the recommended threshold of 0.85
Hypotheses testing
H1, H2, and H4 were tested using OLS regression (see Table 5). As reported in Table 5 (Model 1), the regression model predicting knowledge elaboration was statistically significant (F = 60.366, p < 0.001) and explained 28.6% of the variance in the dependent variable (adjusted R² = 0.286). GenAI acceptance was positively associated with knowledge elaboration (b = 0.352, p < 0.001), and future orientation was also positively associated with knowledge elaboration (b = 0.335, p < 0.001). These results provide support for H1.
Table 5OLS regression results
DV=Knowledge elaboration DV=Service performance
Model 1: direct effect Model 2 moderation effect Model 3: direct effect
b S.E b S.E b S.E
Intercepts 1.307*** 0.307 5.181*** 0.231 1.912*** 0.239
Controls
Education 0.014 0.047 0.024 0.046 -0.067 0.036
Tenure -0.066* 0.029 -0.070* 0.029 0.056* 0.022
Firm age 0.027 0.030 0.024 0.030 -0.012 0.023
Firm size -0.028 0.028 -0.029 0.028 0.021 0.022
Direct effects
GenAI acceptance (GA) 0.352*** 0.036 0.373*** 0.036 0.150*** 0.029
Future orientation (FO) 0.335*** 0.042 0.371*** 0.043 0.363*** 0.033
Knowledge elaboration 0.187*** 0.026
Interaction effects
GA x FO 0.061** 0.019
F value 60.366*** 53.774*** 82.161***
R2 0.291 0.299 0.395
Adjusted R2 0.286 0.294 0.390
N = 890; *p < 0.05, **p < 0.01, ***p < 0.001. Entries are unstandardized regression coefficients
Model 2 introduced the interaction between mean-centered GenAI acceptance and mean-centered future orientation to predict knowledge elaboration. The model explained a substantial proportion of the variance in knowledge elaboration (adjusted R² = 0.294). The interaction term was significant (b = 0.061, p = 0.001), indicating that the positive association between GenAI acceptance and knowledge elaboration varied across levels of future orientation. Both mean-centered GenAI acceptance (b = 0.373, p < 0.001) and mean-centered future orientation (b = 0.371, p < 0.001) were positively associated with knowledge elaboration. Thus, H4 is supported. Figure 2 illustrates the moderating role of future orientation in the association between GenAI acceptance and knowledge elaboration. For visualization, low and high future orientation are plotted at one standard deviation below and above the mean, respectively.
As shown in Table 5 (Model 3), the regression model predicting service performance was statistically significant, explaining 39% of the variance in the dependent variable (adjusted R² = 0.390). Knowledge elaboration was positively associated with service performance (b = 0.187, p < 0.001). Future orientation also remained a significant positive predictor (b = 0.363, p < 0.001). Therefore, H2 is supported.
Using the PROCESS Model 4, we examined whether knowledge elaboration mediated the relationship between GenAI acceptance and service performance while controlling for education, tenure, firm age, and firm size. The model predicting knowledge elaboration was significant, explaining 23.9% of its variance (R² = 0.239). GenAI acceptance was positively associated with knowledge elaboration (b = 0.510, p < 0.001). The difference from the corresponding OLS estimate reflects the inclusion of future orientation as an additional predictor in Model 1 of Table 5. The model predicting service performance was also significant, explaining 31.3% of its variance (R² = 0.313). In the model predicting service performance, both GenAI acceptance (b = 0.284, p < 0.001) and knowledge elaboration (b = 0.261, p < 0.001) were positively associated with service performance. The estimated indirect effect through knowledge elaboration was 0.133, and its 95% bootstrap confidence interval excluded zero (BootLLCI = 0.099, BootULCI = 0.172). The direct association between GenAI acceptance and service performance also remained significant. This pattern was consistent with the proposed partial mediation model and supported H3. However, because knowledge elaboration and service performance were measured at the same wave, the indirect effect should not be interpreted as definitive evidence of temporal or causal mediation.
To test the proposed moderated mediation, we used PROCESS Model 7, in which the interaction between GenAI acceptance and future orientation predicted knowledge elaboration. The model explained 29.9% of the variance in knowledge elaboration (R² = 0.299), and the interaction term was significant (b = 0.061, p = 0.001), indicating that future orientation moderated the association between GenAI acceptance and knowledge elaboration. Conditional effects were examined at relatively low (16th percentile; FO = 5), median (50th percentile; FO = 6), and relatively high (84th percentile; FO = 7) levels of future orientation. The positive association between GenAI acceptance and knowledge elaboration increased from b = 0.324 (SE = 0.037, p < 0.001, 95% CI [0.252, 0.396]) at the 16th percentile to b = 0.385 (SE = 0.037, p < 0.001, 95% CI [0.313, 0.458]) at the 50th percentile and b = 0.446 (SE = 0.046, p < 0.001, 95% CI [0.356, 0.537]) at the 84th percentile. These results support H4.
For service performance, the model, which included knowledge elaboration and control variables, explained 31.3% of the variance (R² = 0.313). The direct association between GenAI acceptance and service performance remained significant (b = 0.284, p < 0.001), and knowledge elaboration was positively associated with service performance (b = 0.261, p < 0.001). These findings suggest that GenAI acceptance is linked to service performance both directly and indirectly through knowledge elaboration. The estimated indirect association between GenAI acceptance and service performance through knowledge elaboration varied across the examined levels of future orientation. The conditional indirect effect was 0.085 at the 16th percentile (FO = 5; BootLLCI = 0.056, BootULCI = 0.115), 0.101 at the 50th percentile (FO = 6; BootLLCI = 0.071, BootULCI = 0.134), and 0.116 at the 84th percentile (FO = 7; BootLLCI = 0.079, BootULCI = 0.160). None of the bootstrap confidence intervals included zero. The index of moderated mediation was 0.016, and its 95% bootstrap confidence interval excluded zero (BootLLCI = 0.003, BootULCI = 0.033). This significant index provided evidence for moderated mediation and supported H5.
Discussion
Discussion of key findings
This study yielded three key findings. First, GenAI acceptance was positively associated with knowledge elaboration, which in turn was positively related to supervisor-rated service performance. The significant indirect effect, together with the remaining direct association between GenAI acceptance and service performance, indicates that knowledge elaboration represents an important, but not exclusive, cognitive pathway through which employees’ acceptance of GenAI is associated with service performance. This finding complements prior research documenting positive productivity and performance outcomes associated with GenAI and employee–AI interaction [7, 23]. The present pattern further suggests that the association between GenAI acceptance and service performance involves employees’ active interpretation, evaluation, refinement, and integration of AI-generated knowledge [27].
Second, the positive association between GenAI acceptance and knowledge elaboration was stronger at higher levels of future orientation. Correspondingly, the estimated indirect association between GenAI acceptance and service performance through knowledge elaboration was also stronger at higher levels of future orientation. This finding indicates that employees with comparable levels of GenAI acceptance do not necessarily engage with its knowledge outputs to the same extent. Employees with stronger future orientation appear more likely to interpret GenAI as a resource for long-term learning, skill development, and career adaptation and therefore to invest more cognitive effort in its outputs. This interpretation is consistent with qualitative evidence that employees proactively reshape their work with GenAI while considering its implications for skill development and career prospects [30]. The present findings extend this research by identifying future orientation as a psychological boundary condition of the GenAI acceptance–knowledge elaboration relationship.
Third, the findings are consistent with the central STS argument that workplace outcomes associated with GenAI reflect the interplay between the technical resource and employees’ psychological and cognitive responses. GenAI acceptance reflects psychological alignment with the technical resource, knowledge elaboration facilitates the cognitive integration of AI-generated knowledge into service practices, and future orientation strengthens this adaptation process. The findings are therefore consistent with research emphasizing the context-dependent and multilevel nature of workplace AI outcomes [18, 24]. They extend this perspective by specifying an employee-level adaptation pathway through which GenAI acceptance is associated with service performance in customer-facing work.
From a psychological perspective, service performance in this study represents the behavioral manifestation of employees’ adaptation to GenAI-enabled work rather than merely a distal organizational outcome. Employees first form psychological evaluations of GenAI, cognitively engage with and elaborate on its outputs, and express this adaptation through context-sensitive service behaviors. This interpretation is consistent with research emphasizing that employees actively regulate cognitive effort, evaluate AI-generated outputs, and adapt their work practices when collaborating with GenAI [27, 30]. Accordingly, the findings advance psychological understanding of human–AI interaction by linking technology-related evaluations, cognitive knowledge processing, and future-oriented individual differences to observable work behavior.
Theoretical contributions
This study makes three main theoretical contributions. First, it refines STS explanations of GenAI-enabled service work by specifying an employee-level adaptation process involving psychological acceptance, cognitive integration, and service performance. Rather than treating GenAI as a technical input with uniform performance consequences, our findings indicate that its value is associated with the alignment between the technical resource and employees’ psychological and cognitive responses. This finding extends STS research by showing how such alignment may unfold in customer-facing service work [18, 24].
Second, this study identifies knowledge elaboration as a theoretically relevant cognitive adaptation pathway consistent with the observed indirect association between GenAI acceptance and employee service performance. Prior research has documented the productivity, creativity, and service-related benefits associated with GenAI and employee–AI interaction [7, 23], but has provided less explanation of how employees transform AI-generated outputs into context-sensitive service actions. The significant indirect association is consistent with the argument that favorable evaluations of GenAI become relevant to service performance partly through employees’ reflection on, evaluation of, refinement of, and integration of its outputs. Because the direct association between GenAI acceptance and service performance remained significant, knowledge elaboration represents an important but nonexclusive explanatory pathway. This finding provides a process-oriented extension of the knowledge-based view [19, 27].
Third, this study identifies future orientation as an individual psychological boundary condition of GenAI-related cognitive adaptation. Employees with comparable levels of GenAI acceptance do not necessarily elaborate on AI-generated knowledge to the same extent. Those with stronger future orientation are more likely to interpret GenAI as a resource for future learning, skill development, and career adaptation and therefore engage more deeply with its outputs. This finding extends future orientation beyond its conventional role as a generalized motivational resource by showing how it conditions the relationship between technological acceptance and cognitive integration in GenAI-enabled work [14, 30].
Practical implications
This study offers several practical implications for service organizations seeking to support employee adaptation to GenAI-enabled work. First, because GenAI acceptance was positively associated with knowledge elaboration, managers should help employees understand the usefulness, ease of use, and manageability of GenAI in specific service tasks. Training should use realistic customer-service scenarios to demonstrate how GenAI can support complaint handling, service recovery, and personalized recommendations. Rather than treating GenAI merely as a tool for automating isolated tasks, firms should embed it into customer-facing workflows in ways that support employees’ interpretation of information, problem solving, and service delivery [1, 10]. Such implementation should augment employee judgment and discretion rather than replace them, particularly because service performance depends heavily on employees’ real-time responsiveness to customer needs [21, 22].
Second, the estimated indirect association between GenAI acceptance and service performance was consistent with mediation through knowledge elaboration. Accordingly, organizations should focus not only on providing access to GenAI but also on helping employees process and apply AI-generated knowledge. Training should teach employees not merely how to obtain GenAI outputs, but how to verify their accuracy, refine their relevance, compare them with organizational policies and professional knowledge, and contextualize them for specific customer needs. Organizations can reinforce these capabilities by developing internal prompt libraries and case repositories and by introducing team debriefings or feedback routines in which employees discuss successful and unsuccessful AI-assisted service cases. These practices can help employees move beyond passive acceptance of AI-generated information and engage in deeper reflection, evaluation, and integration [11–13].
Third, because the positive association between GenAI acceptance and knowledge elaboration was stronger at higher levels of future orientation, a uniform implementation strategy may not be equally appropriate for all employees. Managers could invite highly future-oriented employees to serve as early adopters or peer mentors because they are more likely to view GenAI as a resource for long-term learning, skill development, and career adaptation [14, 30]. Employees with lower future orientation may benefit from more structured support, including coaching, guided practice, pilot programs, and staged implementation. Managers could also connect immediate GenAI-supported tasks with longer-term skill-development and career goals to encourage deeper engagement with AI-generated knowledge. More broadly, successful GenAI implementation should be treated as a sociotechnical change effort in which technological investment is accompanied by employee learning, managerial support, and workflow adaptation [18].
Limitations and future research directions
Notwithstanding its contributions, this study has several limitations that suggest directions for future research. First, although the two-wave, multi-source design reduced common method concerns and temporally separated GenAI acceptance from knowledge elaboration, it does not establish the full temporal sequence implied by the mediation model. Knowledge elaboration and service performance were measured at the same wave, although from different sources. Consequently, the design cannot determine whether knowledge elaboration precedes service performance, whether stronger service performance promotes subsequent knowledge elaboration, or whether the relationship is reciprocal. Future research could use three-wave longitudinal designs, panel models, or experience-sampling methods to establish temporal ordering and examine how psychological acceptance, knowledge elaboration, and service performance evolve over time [39–41]. Accordingly, the significant indirect effect should be interpreted as being consistent with the proposed mediation model rather than as definitive evidence that knowledge elaboration temporally or causally transmits the effect of GenAI acceptance to service performance [42, 43].
Second, this study focused on GenAI acceptance rather than actual GenAI use. Acceptance captures employees’ evaluations of the usefulness, ease of use, and manageability of GenAI, whereas actual use involves behavioral dimensions such as frequency, intensity, duration, and task-specific application. Employees may hold favorable evaluations of GenAI without using it extensively, while frequent use may also result from organizational requirements rather than genuine psychological acceptance. Future research should therefore examine acceptance and actual use simultaneously and distinguish among different forms of GenAI use, such as automation and augmentation, because they may create different cognitive demands and rely on different knowledge-processing mechanisms [25, 44].
Third, the sample covered multiple service industries that may differ in customer-contact intensity, task characteristics, regulatory requirements, and opportunities for GenAI use. Moreover, a substantial proportion of respondents worked in the hotel and catering industry. Although the multi-industry sample provides contextual breadth, the findings may not generalize uniformly across all forms of service work because occupational and organizational contexts can shape employee behavior [45]. For example, knowledge elaboration may play a greater role in customized and knowledge-intensive services than in more standardized service settings. Future research could examine the model across service industries and test whether industry characteristics condition employee adaptation to GenAI.
Fourth, this study focused on employee service performance as the behavioral outcome of adaptation to GenAI-enabled work. However, GenAI may have broader psychological consequences that are not captured by service performance alone. Although GenAI can provide learning, information, and problem-solving resources, it may also increase cognitive demands, uncertainty, concerns about skill obsolescence, and dependence on technological systems. Future research could therefore examine psychological outcomes such as stress, well-being, resilience, thriving, and technology-related anxiety. It could also investigate whether knowledge elaboration and future orientation help employees cope with the psychological demands of human–AI collaboration or, under some conditions, intensify cognitive overload and strain [24, 46].
Conclusion
Drawing on STS theory, this study examined how employees’ psychological and cognitive adaptation to GenAI-enabled service work is associated with service performance. The findings indicate that GenAI acceptance is linked to employee service performance both directly and indirectly through knowledge elaboration and that the estimated indirect association is stronger among employees with higher future orientation. These statistical patterns are consistent with the proposed moderated mediation model, although they should not be interpreted as definitive evidence of causal mediation. Overall, the findings suggest that GenAI-related service outcomes are associated not only with technological availability but also with employees’ psychological acceptance, cognitive engagement with AI-generated knowledge, and future-oriented capacity to adapt to technological change.
Supplementary information
Below is the link to the electronic supplementary material.
References
1.Benbya H Strich F Tamm T Navigating generative artificial intelligence promises and perils for knowledge and creative work J Association Inform Syst 2024 25 1 23 36 10.17705/1jais.00861 Benbya H, Strich F, Tamm T. Navigating generative artificial intelligence promises and perils for knowledge and creative work. J Association Inform Syst. 2024;25(1):23–36. 10.17705/1jais.00861 .
2.Wang S Zhang H Generative AI in international hotel marketing: Impacts on employee creativity and performance Int J Contemp Hospitality Manage 2025 37 8 2601 26 10.1108/IJCHM-10-2024-1595 Wang S, Zhang H. Generative AI in international hotel marketing: Impacts on employee creativity and performance. Int J Contemp Hospitality Manage. 2025;37(8):2601–26. 10.1108/IJCHM-10-2024-1595 .
3.Kanbach DK Heiduk L Blueher G Schreiter M Lahmann A The GenAI is out of the bottle: Generative artificial intelligence from a business model innovation perspective RMS 2024 18 4 1189 220 10.1007/s11846-023-00696-z Kanbach DK, Heiduk L, Blueher G, Schreiter M, Lahmann A. The GenAI is out of the bottle: Generative artificial intelligence from a business model innovation perspective. RMS. 2024;18(4):1189–220. 10.1007/s11846-023-00696-z .
4.Schecter A, Richardson B. (2025). How the role of generative AI shapes perceptions of value in human-AI collaborative work. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (Article 530, pp. 1–15). Association for Computing Machinery. 10.1145/3706598.3713946
5.Quan J, Li Y, Zhang M, Wu Z. Gain or pain: The double-edged sword effect of artificial intelligence adoption on employee service performance. J Organizational Behav. 2025;1–30. 10.1002/job.70023 .
6.Dell’Acqua F McFowland E Mollick III Lifshitz ER Kellogg H Rajendran KC Krayer S Candelon L Lakhani KR Navigating the jagged technological frontier: Field experimental evidence of the effects of artificial intelligence on knowledge worker productivity and quality Organ Sci 2026 37 2 403 23 10.1287/orsc.2025.21838 Dell’Acqua F, McFowland E, Mollick III, Lifshitz ER, Kellogg H, Rajendran KC, Krayer S, Candelon L, F., Lakhani KR. Navigating the jagged technological frontier: Field experimental evidence of the effects of artificial intelligence on knowledge worker productivity and quality. Organ Sci. 2026;37(2):403–23. 10.1287/orsc.2025.21838 .
7.Noy S Zhang W Experimental evidence on the productivity effects of generative artificial intelligence Science 2023 381 6654 187 92 10.1126/science.adh2586 37440646 Noy S, Zhang W. Experimental evidence on the productivity effects of generative artificial intelligence. Science. 2023;381(6654):187–92. 10.1126/science.adh2586 .37440646
8.Zhang Z Dai Y Guo J Zhang H Shen Y Xiao H The double-edged impact of generative artificial intelligence dependence on service innovation in the hospitality industry: A self-regulation perspective Int J Hospitality Manage 2026 132 104392 10.1016/j.ijhm.2025.104392 Zhang Z, Dai Y, Guo J, Zhang H, Shen Y, Xiao H. The double-edged impact of generative artificial intelligence dependence on service innovation in the hospitality industry: A self-regulation perspective. Int J Hospitality Manage. 2026;132:104392. 10.1016/j.ijhm.2025.104392 .
9.Hai S Long T Honora A Japutra A Guo T The dark side of employee-generative AI collaboration in the workplace: An investigation on work alienation and employee expediency Int J Inf Manag 2025 83 102905 10.1016/j.ijinfomgt.2025.102905 Hai S, Long T, Honora A, Japutra A, Guo T. The dark side of employee-generative AI collaboration in the workplace: An investigation on work alienation and employee expediency. Int J Inf Manag. 2025;83:102905. 10.1016/j.ijinfomgt.2025.102905 .
10.Alavi M Leidner DE Mousavi R A knowledge management perspective of generative artificial intelligence J Association Inform Syst 2024 25 1 1 12 10.17705/1jais.00859 Alavi M, Leidner DE, Mousavi R. A knowledge management perspective of generative artificial intelligence. J Association Inform Syst. 2024;25(1):1–12. 10.17705/1jais.00859 .
11.van Ginkel WP Tindale RS van Knippenberg D Knowledge about the distribution of information and group decision making: When and why does it work? Organ Behav Hum Decis Process 2009 108 2 218 29 10.1016/j.obhdp.2008.07.009 van Ginkel WP, Tindale RS, van Knippenberg D. Knowledge about the distribution of information and group decision making: When and why does it work? Organ Behav Hum Decis Process. 2009;108(2):218–29. 10.1016/j.obhdp.2008.07.009 .
12.Ling C-D He W Gong Y Liu W Cho V Does receiving knowledge catalyze creativity? A dyadic-level contingency model of knowledge type and psychological closeness on knowledge elaboration J Organizational Behav 2023 44 9 1436 63 10.1002/job.2741 Ling C-D, He W, Gong Y, Liu W, Cho V. Does receiving knowledge catalyze creativity? A dyadic-level contingency model of knowledge type and psychological closeness on knowledge elaboration. J Organizational Behav. 2023;44(9):1436–63. 10.1002/job.2741 .
13.van Knippenberg D De Dreu CKW Homan AC Work group diversity and group performance: An integrative model and research agenda J Appl Psychol 2004 89 6 1008 22 10.1037/0021-9010.89.6.1008 15584838 van Knippenberg D, De Dreu CKW, Homan AC. Work group diversity and group performance: An integrative model and research agenda. J Appl Psychol. 2004;89(6):1008–22. 10.1037/0021-9010.89.6.1008 .15584838
14.Shipp AJ Edwards JR Lambert LS Conceptualization and measurement of temporal focus: The subjective experience of the past, present, and future Organ Behav Hum Decis Process 2009 110 1 1 22 10.1016/j.obhdp.2009.05.001 Shipp AJ, Edwards JR, Lambert LS. Conceptualization and measurement of temporal focus: The subjective experience of the past, present, and future. Organ Behav Hum Decis Process. 2009;110(1):1–22. 10.1016/j.obhdp.2009.05.001 .
15.Liang X Guo G Shu L Gong Q Luo P Investigating the double-edged sword effect of AI awareness on employee’s service innovative behavior Tour Manag 2022 92 104564 10.1016/j.tourman.2022.104564 Liang X, Guo G, Shu L, Gong Q, Luo P. Investigating the double-edged sword effect of AI awareness on employee’s service innovative behavior. Tour Manag. 2022;92:104564. 10.1016/j.tourman.2022.104564 .
16.Trist EL Bamforth KW Some social and psychological consequences of the longwall method of coal-getting: An examination of the psychological situation and defences of a work group in relation to the social structure and technological content of the work system Hum Relat 1951 4 1 3 38 10.1177/001872675100400101 Trist EL, Bamforth KW. Some social and psychological consequences of the longwall method of coal-getting: An examination of the psychological situation and defences of a work group in relation to the social structure and technological content of the work system. Hum Relat. 1951;4(1):3–38. 10.1177/001872675100400101 .
17.Appelbaum SH Socio-technical systems theory: An intervention strategy for organizational development Manag Decis 1997 35 6 452 63 10.1108/00251749710173823 Appelbaum SH. Socio-technical systems theory: An intervention strategy for organizational development. Manag Decis. 1997;35(6):452–63. 10.1108/00251749710173823 .
18.Makarius EE Mukherjee D Fox JD Fox AK Rising with the machines: A sociotechnical framework for bringing artificial intelligence into the organization J Bus Res 2020 120 262 73 10.1016/j.jbusres.2020.07.045 Makarius EE, Mukherjee D, Fox JD, Fox AK. Rising with the machines: A sociotechnical framework for bringing artificial intelligence into the organization. J Bus Res. 2020;120:262–73. 10.1016/j.jbusres.2020.07.045 .
19.Grant RM Toward a knowledge-based theory of the firm Strateg Manag J 1996 17 S2 109 22 10.1002/smj.4250171110 Grant RM. Toward a knowledge-based theory of the firm. Strateg Manag J. 1996;17(S2):109–22. 10.1002/smj.4250171110 .
20.Podsakoff PM Podsakoff NP Williams LJ Huang C Yang J Common method bias: It’s bad, it’s complex, it’s widespread, and it’s not easy to fix Annual Rev Organizational Psychol Organizational Behav 2024 11 1 17 61 10.1146/annurev-orgpsych-110721-040030 Podsakoff PM, Podsakoff NP, Williams LJ, Huang C, Yang J. Common method bias: It’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annual Rev Organizational Psychol Organizational Behav. 2024;11(1):17–61. 10.1146/annurev-orgpsych-110721-040030 .
21.Liao H Chuang A A multilevel investigation of factors influencing employee service performance and customer outcomes Acad Manag J 2004 47 1 41 58 10.5465/20159559 Liao H, Chuang A. A multilevel investigation of factors influencing employee service performance and customer outcomes. Acad Manag J. 2004;47(1):41–58. 10.5465/20159559 .
22.Hur WM Shin Y Moon TW Linking motivation, emotional labor, and service performance from a self-determination perspective J Service Res 2022 25 2 227 41 10.1177/1094670520975204 Hur WM, Shin Y, Moon TW. Linking motivation, emotional labor, and service performance from a self-determination perspective. J Service Res. 2022;25(2):227–41. 10.1177/1094670520975204 .
23.Yang S Guo Y Chen Y Li Y Zhou S Jiang H Understanding service performance of frontline employees from employee–AI interaction perspective J Service Manage 2025 36 3 448 75 10.1108/JOSM-10-2023-0425 Yang S, Guo Y, Chen Y, Li Y, Zhou S, Jiang H. Understanding service performance of frontline employees from employee–AI interaction perspective. J Service Manage. 2025;36(3):448–75. 10.1108/JOSM-10-2023-0425 .
24.Bankins S, Weaver A, Marrone M, Restubog SLD, Woo SE. Artificial intelligence at work: A phenomenon-based interdisciplinary review and groundwork for multilevel scholarship. Inf Manag. 2026;63(2). 10.1016/j.im.2025.104272 .
25.Davis FD Perceived usefulness, perceived ease of use, and user acceptance of information technology MIS Q 1989 13 3 319 40 10.2307/249008 Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 1989;13(3):319–40. 10.2307/249008 .
26.Venkatesh V Bala H Technology acceptance model 3 and a research agenda on interventions Decis Sci 2008 39 2 273 315 10.1111/j.1540-5915.2008.00192.x Venkatesh V, Bala H. Technology acceptance model 3 and a research agenda on interventions. Decis Sci. 2008;39(2):273–315. 10.1111/j.1540-5915.2008.00192.x .
27.Memmert L Soroko D Bittner E From effort reduction to effort management: An expectancy theory perspective on professionals’ work practices with generative AI Bus Inform Syst Eng 2025 67 615 35 10.1007/s12599-025-00960-4 Memmert L, Soroko D, Bittner E. From effort reduction to effort management: An expectancy theory perspective on professionals’ work practices with generative AI. Bus Inform Syst Eng. 2025;67:615–35. 10.1007/s12599-025-00960-4 .
28.Setia P Venkatesh V Joglekar S Leveraging digital technologies: How information quality leads to localized capabilities and customer service performance MIS Q 2013 37 2 565 90 10.25300/MISQ/2013/37.2.11 Setia P, Venkatesh V, Joglekar S. Leveraging digital technologies: How information quality leads to localized capabilities and customer service performance. MIS Q. 2013;37(2):565–90. 10.25300/MISQ/2013/37.2.11 .
29.Praskova A Johnston L The role of future orientation and negative career feedback in career agency and career success in Australian adults J Career Assess 2021 29 3 463 85 10.1177/1069072720980174 Praskova A, Johnston L. The role of future orientation and negative career feedback in career agency and career success in Australian adults. J Career Assess. 2021;29(3):463–85. 10.1177/1069072720980174 .
30.Mayer A-S Baygi RM Buwalda R Generation AI: Job crafting by entry-level professionals in the age of generative AI Bus Inform Syst Eng 2025 67 595 613 10.1007/s12599-025-00959-x Mayer A-S, Baygi RM, Buwalda R. Generation AI: Job crafting by entry-level professionals in the age of generative AI. Bus Inform Syst Eng. 2025;67:595–613. 10.1007/s12599-025-00959-x .
31.MacKenzie SB Podsakoff PM Common method bias in marketing: Causes, mechanisms, and procedural remedies J Retail 2012 88 4 542 55 10.1016/j.jretai.2012.08.001 MacKenzie SB, Podsakoff PM. Common method bias in marketing: Causes, mechanisms, and procedural remedies. J Retail. 2012;88(4):542–55. 10.1016/j.jretai.2012.08.001 .
32.Goodman JS Blum TC Assessing the non-random sampling effects of subject attrition in longitudinal research J Manag 1996 22 4 627 52 10.1177/014920639602200405 Goodman JS, Blum TC. Assessing the non-random sampling effects of subject attrition in longitudinal research. J Manag. 1996;22(4):627–52. 10.1177/014920639602200405 .
33.Brislin RW. Translation and content analysis of oral and written material. In: Triandis HC, Berry JW, editors. Handbook of cross-cultural psychology. Volume 2. Allyn & Bacon; 1980. pp. 349–444.
34.Singh K Chatterjee S Mariani M Applications of generative AI and future organizational performance: The mediating role of explorative and exploitative innovation and the moderating role of ethical dilemmas and environmental dynamism Technovation 2024 133 103021 10.1016/j.technovation.2024.103021 Singh K, Chatterjee S, Mariani M. Applications of generative AI and future organizational performance: The mediating role of explorative and exploitative innovation and the moderating role of ethical dilemmas and environmental dynamism. Technovation. 2024;133:103021. 10.1016/j.technovation.2024.103021 .
35.Khan S Mehmood S Khan SU Navigating innovation in the age of AI: How generative AI and innovation influence organizational performance in the manufacturing sector J Manuf Technol Manage 2025 36 3 597 620 10.1108/JMTM-06-2024-0302 Khan S, Mehmood S, Khan SU. Navigating innovation in the age of AI: How generative AI and innovation influence organizational performance in the manufacturing sector. J Manuf Technol Manage. 2025;36(3):597–620. 10.1108/JMTM-06-2024-0302 .
36.Lu Y Yang MM Mitchell R Boyle B Gu J Algorithmic reliability at the helm: Investigating the relationship between experienced algorithmic reliability, trust, and work engagement in the gig economy Hum Resour Manag 2026 65 2 493 509 10.1002/hrm.70033 Lu Y, Yang MM, Mitchell R, Boyle B, Gu J. Algorithmic reliability at the helm: Investigating the relationship between experienced algorithmic reliability, trust, and work engagement in the gig economy. Hum Resour Manag. 2026;65(2):493–509. 10.1002/hrm.70033 .
37.Hayes AF. Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. 2nd ed. Guilford Press; 2018.
38.Voorhees CM Brady MK Calantone R Ramirez E Discriminant validity testing in marketing: An analysis, causes for concern, and proposed remedies J Acad Mark Sci 2016 44 1 119 34 10.1007/s11747-015-0455-4 Voorhees CM, Brady MK, Calantone R, Ramirez E. Discriminant validity testing in marketing: An analysis, causes for concern, and proposed remedies. J Acad Mark Sci. 2016;44(1):119–34. 10.1007/s11747-015-0455-4 .
39.Allen TD, French KA, Braun MT, Fletcher K. The passage of time in work-family research: Toward a more dynamic perspective. J Vocat Behav 110(Part B. 2019;245–57. 10.1016/j.jvb.2018.11.013 .
40.Ployhart RE Vandenberg RJ Longitudinal research: The theory, design, and analysis of change J Manag 2010 36 1 94 120 10.1177/0149206309352110 Ployhart RE, Vandenberg RJ. Longitudinal research: The theory, design, and analysis of change. J Manag. 2010;36(1):94–120. 10.1177/0149206309352110 .
41.Li X, Qing L, Wei H, Wei J. Artificial intelligence and innovation ambidexterity in advanced manufacturing firms: a knowledge-management perspective. J Manuf Technol Manage. 2026;1–26. 10.1108/JMTM-06-2025-0489 .
42.Cole DA Maxwell SE Testing Mediational Models With Longitudinal Data: Questions and Tips in the Use of Structural Equation Modeling J Abnorm Psychol 2003 112 4 558 77 10.1037/0021-843X.112.4.558 14674869 Cole DA, Maxwell SE. Testing Mediational Models With Longitudinal Data: Questions and Tips in the Use of Structural Equation Modeling. J Abnorm Psychol. 2003;112(4):558–77. 10.1037/0021-843X.112.4.558 .14674869
43.Maxwell SE Cole DA Bias in cross-sectional analyses of longitudinal mediation Psychol Methods 2007 12 1 23 44 10.1037/1082-989X.12.1.23 17402810 Maxwell SE, Cole DA. Bias in cross-sectional analyses of longitudinal mediation. Psychol Methods. 2007;12(1):23–44. 10.1037/1082-989X.12.1.23 .17402810
44.Guo M Gu M Huo B The impacts of automation and augmentation AI use on physicians’ performance: An ambidextrous perspective Int J Oper Prod Manage 2025 45 1 114 51 10.1108/IJOPM-06-2023-0509 Guo M, Gu M, Huo B. The impacts of automation and augmentation AI use on physicians’ performance: An ambidextrous perspective. Int J Oper Prod Manage. 2025;45(1):114–51. 10.1108/IJOPM-06-2023-0509 .
45.Johns G The essential impact of context on organizational behavior Acad Manage Rev 2006 31 2 386 408 10.5465/amr.2006.20208687 Johns G. The essential impact of context on organizational behavior. Acad Manage Rev. 2006;31(2):386–408. 10.5465/amr.2006.20208687 .
46.Hermann E Puntoni S Morewedge CK GenAI and the psychology of work Trends Cogn Sci 2025 29 9 802 13 10.1016/j.tics.2025.04.009 40345947 Hermann E, Puntoni S, Morewedge CK. GenAI and the psychology of work. Trends Cogn Sci. 2025;29(9):802–13. 10.1016/j.tics.2025.04.009 .40345947
Come citare questo studio
Li X, Ou Y, Wang J. (2026). Psychological adaptation to generative AI in service work: knowledge elaboration and future orientation.. BMC psychology. DOI: 10.1186/s40359-026-05477-3
Consulta la pubblicazione originale ↗