Review Reports - The Impact of Digital Leadership on Employee Resilience: The Mediating Roles of Work Gamification and Workplace Mindfulness and the Moderating Role of AI Anxiety

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear author(s),

This manuscript integrates digital leadership, work gamification and artificial intelligence anxiety into the research framework of employees' resilience, and responds to the research needs of employees ' occupational health in the context of digital transformation. This paper incorporates digital leadership, work gamification, and artificial intelligence anxiety into the research framework of employee resilience, and responds to the research needs of employee occupational health in the context of digital transformation. People-oriented, fully reveals the interaction between enterprise organizational environment, employee individual cognition and employee behavior. At the same time, the current version has several conceptual, methodological, hypothetical derivation rigor and editorial weaknesses that need to be addressed before the manuscript can be considered for publication. My comments below are intended to help strengthen this paper.

First of all, the expression of the research question is not clear enough. The introduction does not define the applicable scenarios and the boundary of the research object. Readers cannot understand whether this study is aimed at all enterprises under the digital background, or at enterprises in the middle of digital transformation and with high AI anxiety of employees. At the same time, the research gap is not focused on the lack of systematic correlation between digital leadership and employees ' psychological resilience, and the explanation of the causes and practical consequences of the gap is simple. Instead, the secondary gap between leadership behavior and employees ' technical anxiety is overemphasized, and there is conceptual confusion.

Second, the use of Social Cognitive Theory is somewhat simple. Based on Bandura 's social cognitive theory, the manuscript should pay attention to the dynamic interaction among environment, individual cognition and behavior. The paper defines work gamification as institutional environmental factors, workplace mindfulness as employees ' individual cognitive factors, and employees' resilience as behavioral results. There is no theoretical explanation for the interaction of the three, failure to explain the theoretical gap. It only analyzes the one-way path from digital leadership to the environment or individuals to psychological resilience in isolation, ignoring the core connotation of individual cognitive reaction to the environment, behavior shaping environment and cognition in social cognitive theory. It can be supplemented to explain whether workplace mindfulness will affect employees ' perception and participation in work gamification, and whether employees' resilience will reversely strengthen the implementation effect of digital leadership or employees' mindfulness level.

Third, the relationship between the core variables are not sorted out and the logic is not strong. The paper takes digital leadership and employee resilience as the core variables, but the literature on the core variables is insufficient, and the dimensions, measurement methods and related research of the variables are not sorted out. The paper takes work gamification and AI anxiety as mediating or moderating variables, but does not sort out any related research on these variables in the introduction, which leads to the lack of literature support for the rationality of incorporating these variables into the model in the future. Readers cannot understand why these variables can become mediating or moderating variables between digital leadership and employee resilience. If the paper can more abundantly introduce the theoretical and practical basis of variable selection, it will be easier for readers to understand the research content of the paper.

Fourth, the connection between the definition and measurement of employees' resilience. In this paper, employee resilience is defined as the ability of employees to adapt to the environment, cope with challenges and achieve self-growth with the active support of the organization. However, the scale of Näswall et al.is used in the measurement. The items of the scale mainly focus on the individual 's crisis response and feedback processing ability. It is not known whether it can reflect the core defining element of organizational support, which affects the logic of the relationship between digital leadership (leadership behavior at the organizational level) and employee resilience. If resilience is only individual ability, the influence path of digital leadership needs to emphasize the cultivation of individual ability rather than the shaping of the organizational environment mentioned in the paper. The logical basis of the variable relationship is thus weakened. I suggest that the author find a scale that is more suitable for the organizational environment, or explain the usability of the scale.

Fifth, the matching problem between the dimension of artificial intelligence anxiety and the relationship between variables. This paper quotes the research of Li Jing and Huang Jinsong, and divides AI anxiety into two dimensions : AI learning anxiety and work substitution anxiety. The influence logic of the two dimensions on employees is different : AI learning anxiety originates from their own lack of ability, which may encourage employees to take the initiative to use work gamification and enhance mindfulness to make up for the ability gap ; job substitution anxiety stems from job safety concerns, which may lead to employees' defensive psychology and reduce their willingness to use organizational resources. The paper regards artificial intelligence anxiety as a single overall variable, does not distinguish the difference between the two, and analyzes it as a positive adjustment variable, which may lead to the inherent logical contradiction of the adjustment effect. I suggest that the authors discuss the two dimensions separately to better reflect the actual results and give more scientific and uncontroversial research results, rather than confusing the two dimensions that may have opposite effects.

Sixth, the explanation of the direct effect is somewhat exaggerated, and the analysis of the mediating effect lacks the discussion of the interaction effect. From the perspective of the interaction model of social cognitive theory, work gamification (environment) is likely to affect workplace mindfulness (individual). However, taking the two as parallel mediating variables, ignoring the potential relationship between the two, and taking it as an independent path analysis may lead to the superposition or masking of mediating effects and affect the accuracy of the model. In the descriptive statistics, it is found that the correlation coefficient between work gamification and workplace mindfulness is as high as 0.565 ( p < 0.01 ), indicating that there is a strong positive correlation between the two. It is suggested to supplement the multicollinearity test, so as to maintain the integrity of the variable relationship.

Seventh, we should expand the discussion on the non-significant results of the mediating role of AI anxiety regulating work gamification between digital leadership and employees' resilience. Because artificial intelligence anxiety is presented as a central explanation mechanism, its lack of significant mediating role deserves more adequate treatment. In view of the theoretical importance given to this construct in the manuscript, this is not sufficient. The authors should consider alternative explanations more carefully, including measurement problems, the limited specificity of AI anxiety measurement, and potential paths that may affect the main effect.

Finally, there are several editing and formatting problems that need to be corrected. These problems include the table cross-page does not set the continuation table, the chart position, the remaining template language, the reference format is not uniform and so on. While these issues are relatively easy to resolve, they give the impression that manuscripts are not yet fully prepared for submission in the final journal format.

In general, I believe that the manuscript solves a meaningful research problem, focusing on the research on the impact of digital leadership on employee resilience in the context of digitalization, and may make useful contributions in practice. However, the current version requires a series of revisions to clarify the measurement dimension of the variables, improve the logic of the hypothesis discussion, make the conclusions and results more closely consistent, and improve the overall presentation quality of the paper.

Best regards.

Author Response

REVISION NOTES

Response to Reviewer 1 Comments

Summary

Thank you very much for taking the time to review this manuscript. Please find our detailed responses below. In the revised files, the content-related modifications are highlighted in blue, while the revisions made to improve the accuracy of the English language are shown using the Track Changes mode.

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[First of all, the expression of the research question is not clear enough. The introduction does not define the applicable scenarios and the boundary of the research object. Readers cannot understand whether this study is aimed at all enterprises under the digital background, or at enterprises in the middle of digital transformation and with high AI anxiety of employees. At the same time, the research gap is not focused on the lack of systematic correlation between digital leadership and employees ' psychological resilience, and the explanation of the causes and practical consequences of the gap is simple. Instead, the secondary gap between leadership behavior and employees ' technical anxiety is overemphasized, and there is conceptual confusion.]

Response 1：Thank you very much for your valuable comments. We overlooked the research boundaries in our study and failed to conduct a comprehensive analysis of the causes and consequences of the research gaps. As the core variables of this article involve digital leadership and employee resilience, as well as work gamification, workplace mindfulness, and AI anxiety, all of which are highly related to the level of digitalization, this study focuses on enterprises during the digital transformation period. We have also clearly stated the specific purpose of this research in the abstract and the text. In addition, we have supplemented the reasons for the research gaps from the perspectives of environmental development and enterprise nature, and explained the consequences of not conducting the research by combining the strong impact of the digital environment and the paradox of digital transformation.

The additions, deletions and reorganizations of the original content can be seen on section abstract page 1 lines 13-14, section introduction page 2 lines 54-56, section introduction pages 2-3 lines 88-99, and section introduction page 3 lines 102-114 of the revised manuscript.

“[……many enterprises are in the midst of digital transformation, making it increasingly important to safeguard employees' occupational health.(abstract)

This study will focus on enterprises during the period of digital transformation, exploring the relationship between digital leadership and employee resilience.(introduction)

The reason why few studies have combined digital leadership with employee resilience is as follows: Firstly, the recent intensification of digitalization in enterprises has led to digital capabilities not only existing in large enterprises but also being valued by small and medium-sized enterprises, laying a solid foundation for further research by scholars. Secondly, due to the profit-oriented nature of most enterprises, existing research is mostly limited to the organizational level. The primary purpose of introducing digital leadership is to drive the correct transformation of enterprises and consolidate their market position, which makes it easier to overlook the micro-level of individual employees. However, the successful development of an enterprise cannot be achieved without the efforts of its employees. If the relationship between digital leadership and employee resilience is ignored, when digitalization impacts employee resilience, leaders will lack a useful theoretical framework to intervene.(introduction)

Therefore, in the context of a working environment where AI technology is increasingly being utilized, it is of utmost importance to explore how digital leadership can effectively enhance employee resilience and to investigate the role that AI anxiety plays in this process.(introduction)]”

Comment 2：[Second, the use of Social Cognitive Theory is somewhat simple. Based on Bandura 's social cognitive theory, the manuscript should pay attention to the dynamic interaction among environment, individual cognition and behavior. The paper defines work gamification as institutional environmental factors, workplace mindfulness as employees ' individual cognitive factors, and employees' resilience as behavioral results. There is no theoretical explanation for the interaction of the three, failure to explain the theoretical gap. It only analyzes the one-way path from digital leadership to the environment or individuals to psychological resilience in isolation, ignoring the core connotation of individual cognitive reaction to the environment, behavior shaping environment and cognition in social cognitive theory. It can be supplemented to explain whether workplace mindfulness will affect employees ' perception and participation in work gamification, and whether employees' resilience will reversely strengthen the implementation effect of digital leadership or employees' mindfulness level.]

Response 2：Thank you very much for your valuable comments. Based on the findings of this study, we admit that the application of the social cognitive theory was not comprehensive. Therefore, we only used the single path of "environment → individual → behavior" in the social cognitive theory and did not apply the triadic interaction path. We have revised the statements about the triadic interaction in the text. As the model in this study involves the moderating effect of AI anxiety, we have also introduced the conservation of resources theory to explain the moderating effect of AI anxiety and explored the mediating role of individual psychological resources and organizational resources from the perspective of resources. We have also inserted phrases such as "According to the conservation of resources theory" in the original text. Regarding the triadic interaction you mentioned, we will incorporate it into Section 5.4 and hope to further explore these variables in the future.

You can find these changes on section introduction page 3 lines 115-119, section introduction pages 3-4 lines 134-144, and section 5.4 page 22 lines 823-841 of the revised manuscript.

“[To address these shortcomings, the objective of this study is to systematically reveal the mechanism and boundary conditions through which digital leadership influences employee resilience, based on the unidirectional transmission path of "environment → individual → behavior" in Social Cognitive Theory, while also incorporating the Conservation Of Resources Theory .(introduction)

According to the Conservation Of Resources Theory, resources are the driving force of human behavior and psychological states. When resources are at risk of being lost, people will experience stress and tension. AI anxiety is the negative emotion that employees feel when they are worried about losing their jobs due to the impact of AI. On the one hand, employees with high levels of AI anxiety will feel threatened when facing the digital leadership promoted by the enterprise, thus being immersed in tension and finding it difficult to engage in the learning of digital technologies. On the other hand, when the level of AI anxiety is high, employees will mobilize new resources to avoid losing their own resources and follow the organization to mobilize mindfulness resources to enhance their resilience. The manifestation of employee resilience is the result of resource mobilization.(introduction)

Third, this study only explored the relationship between digital leadership and employee resilience from the perspective of "environment → individual → behavior" based on the single path of social cognitive theory. However, it lacks the investigation of the possible different relationships among work gamification, workplace mindfulness, and employee resilience based on the triadic interaction foundation of social cognitive theory. Moreover, due to the limitation of cross-sectional data, this study failed to deeply explore the "individual behavior → environment" path, that is, whether employee resilience would inversely affect work gamification and digital leadership, nor did it delve into the "behavior → individual → environment" path and whether employee resilience would inversely affect the level of mindfulness to promote changes in digital leadership. Future research could adopt panel data for long-term tracking to obtain more accurate research data and deeply analyze whether there are potential mediating paths between work gamification and workplace mindfulness, as well as the relationship between employee resilience and individual and environmental factors.(5.4)]”

Comment 3：[Third, the relationship between the core variables are not sorted out and the logic is not strong. The paper takes digital leadership and employee resilience as the core variables, but the literature on the core variables is insufficient, and the dimensions, measurement methods and related research of the variables are not sorted out. The paper takes work gamification and AI anxiety as mediating or moderating variables, but does not sort out any related research on these variables in the introduction, which leads to the lack of literature support for the rationality of incorporating these variables into the model in the future. Readers cannot understand why these variables can become mediating or moderating variables between digital leadership and employee resilience. If the paper can more abundantly introduce the theoretical and practical basis of variable selection, it will be easier for readers to understand the research content of the paper.]

Response 3：Thank you very much for your suggestions for improvement. We have supplemented the preliminary research on the five core variables, their dimensions and measurement methods, and included the literature of key scholars in this field. We have also explained the rationality of incorporating workplace mindfulness and work gamification as mediating variables and AI anxiety as a moderating variable from the perspectives of other studies and resource conservation theory, respectively.

You can see these changes on section introduction page 2 lines 69-72, section 2.1 page 4 lines 165-173, section 2.1 pages 4-5 lines 184-188, section 2.2 page 5 lines 212-215, section 2.2 page 5 lines 223-224, section 2.2 page 5 lines 227-229, section 2.3 page 6 lines 279-282, section 2.4 page 7 lines 322-323, section 2.4 page 8 lines 339-346, section introduction page 3 lines 119-123, and section introduction pages 3-4 lines 134-144 of the revised manuscript.

“[Employee resilience in the workplace is a core issue in the field of organizational behavior. In terms of measuring employee resilience, there has been a shift from the initial incorrect use of psychological resilience scales to the later accurate application of employee resilience scales.(introduction)

In the digital age, the changes and development of leadership have attracted considerable attention from scholars in recent years. Kane et al. were the first to study how digital disruption has transformed business and leadership, and put forward new requirements for traditional leadership skills (Kane G. C. et al., 2019). Contreras et al. proposed that in the special environment of the pandemic, an effective leadership is needed to guide employees in using electronic technology to enhance the adaptability of enterprises (Contreras Francoise et al., 2020). As digital technology gradually shifts from a technical support role to a key driver of organizational change, the concept and application of digital leadership have also been refined.(2.1)

The existing measurement tools for digital leadership are mainly developed based on Western scenarios. Some use unidimensional scales for measurement, while others use multidimensional scales. Roman et al. divided digital leadership into six dimensions: communication, social, flexibility, team, and trust (Roman, A. V. et al., 2019).(2.1)

Nick Pelling defines gamification as applying game elements to the learning process in any field. Deterding, S. et al. propose that it is the use of game elements in non-gamified domains. Subsequently, gamification has been widely applied in various scenarios (Deterding, S. et al., 2011) ,laying the foundation for the research on work gamification.(2.2)

……the measurement of the degree of gamification within an enterprise’s internal work is usually determined by the number of gamification elements.(2.2)

Integrating game elements into work systems requires the integration of multiple fields such as computer science and game design, and it is not only dependent on relevant digital technologies……(2.2)

In terms of measurement, early measures of workplace mindfulness mostly utilized single-dimensional scales. In recent years, Zheng, X. M. et al. developed a three-dimensional scale for workplace mindfulness based on work scenarios, which has also made the measurement results more accurate.(2.3)

Li, J., and Huang, J. S. hold that AI anxiety is the fear and concern of individuals about losing control due to the rapid development of AI technology(Li, J., &Huang, J. S.,2020).(2.4)

When measuring AI anxiety, multi-dimensional scales are generally used for measurement, covering multiple dimensions to comprehensively measure expected anxiety and annihilation anxiety. Li, J., and Huang, J. S. first comprehensively constructed a theoretical framework for AI anxiety, dividing it into eight core dimensions including privacy invasion anxiety, algorithm bias anxiety, job replacement anxiety, and technology learning anxiety (Li, J., & Huang, J. S., 2020). Wang, Y.-M. et al. further developed a universal AI anxiety scale (AIAS), making subsequent measurements more applicable (Wang, Y.-M. et al., 2022).(2.4)

As a novel work system, work gamification can stimulate employees' positive emotions and help them accumulate psychological resources to adapt to significant environmental changes (Decha Dechawatanapaisal, 2025); Taşkan Burcu et al. found that workplace mindfulness can mediate the relationship between uncertainty and adaptive performance, effectively linking the environment and behavior (Taşkan Burcu et al., 2025).(introduction)

Contreras Francoise,Baykal Elif & Abid Ghulam.(2020).E-Leadership and Teleworking in Times of COVID-19 and Beyond: What We Know and Where Do We Go.Frontiers in Psychology,11,590271-590271.https://doi.org/10.3389/FPSYG.2020.590271.

Deterding, S., Dixon, D., Khaled, R., & Nacke, L. (2011). From game design elements to gamefulness: Defining "gamification". Proceedings of the 15th International Academic MindTrek Conference, 9–15.

Decha Dechawatanapaisal.(2025).Influence of positive affect on readiness for change following the pandemic: a moderated mediation study.Journal of Organizational Change Management,38(4),746-763.https://doi.org/10.1108/JOCM-04-2024-0232.

Kane, G. C., Phillips, A. N., Copulsky, J., & Andrus, G. (2019). How digital leadership is different. MIT Sloan Management Review, 60(3), 34-39.

Taşkan Burcu,Junça-Silva Ana & Caetano António.(2025).The role of mindfulness in the relationship between uncertainty and adaptive performance via affect: a within-person approach.Management Research Review,48(3),489-505.https://doi.org/10.1108/MRR-11-2023-0874.]”

Comment 4：[Fourth, the connection between the definition and measurement of employees' resilience. In this paper, employee resilience is defined as the ability of employees to adapt to the environment, cope with challenges and achieve self-growth with the active support of the organization. However, the scale of Näswall et al.is used in the measurement. The items of the scale mainly focus on the individual 's crisis response and feedback processing ability. It is not known whether it can reflect the core defining element of organizational support, which affects the logic of the relationship between digital leadership (leadership behavior at the organizational level) and employee resilience. If resilience is only individual ability, the influence path of digital leadership needs to emphasize the cultivation of individual ability rather than the shaping of the organizational environment mentioned in the paper. The logical basis of the variable relationship is thus weakened. I suggest that the author find a scale that is more suitable for the organizational environment, or explain the usability of the scale.]

Response 4：Thank you for your valuable suggestions on the definition and scale selection of employee resilience. Our full definition of employee resilience still focuses on an individual's response to crises and their own adaptability, which is consistent with the measurement of employee resilience using the Näswall et al. scale. In our research, we set the organizational environment (digital leadership, work gamification), that is, organizational support, as a preceding variable and proposed that the organizational environment would have an impact on employee resilience. The original text of this study confused the definition of employee resilience with the preceding factor. We have now revised the original text. The phrase "with the help and support of the organization" has been deleted.

You can see these changes on section 2.3 page 7 lines 297-298 of the revised manuscript.

“[Employee resilience refers to the ability of employees to adapt, cope with and grow in the face of constantly changing environments（Tonkin et al., 2018）. (2.3)]”

Comment 5：[Fifth, the matching problem between the dimension of artificial intelligence anxiety and the relationship between variables. This paper quotes the research of Li Jing and Huang Jinsong, and divides AI anxiety into two dimensions : AI learning anxiety and work substitution anxiety. The influence logic of the two dimensions on employees is different : AI learning anxiety originates from their own lack of ability, which may encourage employees to take the initiative to use work gamification and enhance mindfulness to make up for the ability gap ; job substitution anxiety stems from job safety concerns, which may lead to employees' defensive psychology and reduce their willingness to use organizational resources. The paper regards artificial intelligence anxiety as a single overall variable, does not distinguish the difference between the two, and analyzes it as a positive adjustment variable, which may lead to the inherent logical contradiction of the adjustment effect. I suggest that the authors discuss the two dimensions separately to better reflect the actual results and give more scientific and uncontroversial research results, rather than confusing the two dimensions that may have opposite effects.]

Response 5：Thank you for your valuable comments on the selection of the AI Anxiety Scale. Although there are subtle differences between AI replacement anxiety and AI learning anxiety, including AI anxiety can explore the impact of AI anxiety from a macro perspective. First, from the perspective of the resource conservation theory, both AI learning anxiety and AI replacement anxiety can make employees perceive the threat of resource loss. AI replacement anxiety is the perception of the threat to work resources, while AI learning anxiety is the perception of the threat to time and ability resources. Therefore, we believe that the perception and utilization of environmental resources and personal resources are the same. Secondly, we used SPSS to include all six items of AI anxiety in the exploratory factor analysis and extracted factors using the principal component analysis method. The results showed that the KMO value was 0.877, and the Bartlett's sphericity test was significant (approximate chi-square = 923.116, df = 15, p < 0.001). The unrotated factor analysis only extracted one common factor with an eigenvalue greater than 1, with an eigenvalue of 3.838 and a cumulative variance explanation rate of 63.974%. The factor loadings of each item ranged from 0.687 to 0.866, all greater than the critical value of 0.5. The above results indicate that the six items of AI anxiety have a clear one-dimensional structure. Finally, according to the research results of Li Jiang et al., there is no contradiction in the direction of the moderating effect. It may be because the scales selected are different, two dimensions in this study can be combined. If the dimension division is changed to different scales, new scales need to be selected for data analysis. The future research direction of AI replacement anxiety should be the same as that of Li Jiang et al., to refine the dimensions of AI anxiety and select different scales for more rigorous analysis. Therefore, we added the content of EFA in the original text and also added the possible shortcomings of this study in 5.4.

Li Jiang,Xuli Wang,Yuguang Xie,Junhong Zhu & Dongxiao Gu.(2026).Substitution or augmentation? How organisational AI adoption approaches influence the willingness to stay of caregivers.Computers in Human Behavior,174,108801-108801.https://doi.org/10.1016/J.CHB.2025.108801.

You can see this change on section 3.2 pages 11-12 lines 468-471 and section 5.4 page 22 lines 829-834 of the revised manuscript. Once again, thank you for your suggestions, which have given us a direction for research and improvement.

“[The results of exploratory factor analysis indicated that the six items of AI anxiety extracted one common factor (KMO = 0.877, cumulative variance explained rate was 63.974%), with item loadings ranging from 0.687 to 0.866, suggesting that it is feasible to treat AI anxiety as a whole.(3.2)

Finally, there are deficiencies in the measurement of AI anxiety in this paper. This study selected AI anxiety, which includes AI learning anxiety and job replacement anxiety. However, there may be subtle differences between the two. Future research should select two different scales for AI learning anxiety and job replacement anxiety for study, to analyze the relationship between different dimensions of AI anxiety and different variables from a more microscopic perspective, making the research results more accurate.(5.4)]”

Comment 6：[Sixth, the explanation of the direct effect is somewhat exaggerated, and the analysis of the mediating effect lacks the discussion of the interaction effect. From the perspective of the interaction model of social cognitive theory, work gamification (environment) is likely to affect workplace mindfulness (individual). However, taking the two as parallel mediating variables, ignoring the potential relationship between the two, and taking it as an independent path analysis may lead to the superposition or masking of mediating effects and affect the accuracy of the model. In the descriptive statistics, it is found that the correlation coefficient between work gamification and workplace mindfulness is as high as 0.565 ( p < 0.01 ), indicating that there is a strong positive correlation between the two. It is suggested to supplement the multicollinearity test, so as to maintain the integrity of the variable relationship.]

Response 6：Thank you for your valuable suggestions on the interpretation of the direct effect and the relationships among variables. In this study, we exaggerated the interpretation of the direct effect and repeatedly used terms such as "significantly enhanced" and "directly enhanced". These have now been revised to "significantly positively predicted" and "positively promoted". Secondly, in response to the possibility that work gamification might affect workplace mindfulness and thus form a chain mediation, we have supplemented the theoretical analysis of this study, additional experiments, and conducted a multicollinearity test. The results show that there is no possibility of chain mediation and no multicollinearity problem. We also proposed in 5.4 that future research should conduct in-depth analysis of the mediating path between work gamification and workplace mindfulness through longitudinal data.

Variable	Tolerance	VIF
DL	0.714	1.401
WG	0.520	1.921
AIA	0.466	2.146
WM	0.457	2.188

You can find these changes on section abstract page 1 lines 22-23, section introduction page 3 line 130, section 2.3 page 9 lines 403-414, section 4.3 page 13 lines 506-510, section 4.5 page 17 lines 603-609, and section 5.4 page 22 lines 825-828 of the revised manuscript.

“[Survey data from 293 employees show that digital leadership significantly and positively predicts employee resilience.(abstract)

digital leadership not only positively promote employee resilience(introduction)

This study incorporates both work gamification and workplace mindfulness simultaneously into the mediating path, without establishing a chain mediation through which digital leadership influences employee resilience via work gamification and workplace mindfulness. The main reasons are as follows: Firstly, the data collected in this study is cross-sectional, which cannot effectively capture the chain mediating relationship between gamification and mindfulness. Chain mediation requires a clear temporal sequence, but the questionnaire data is difficult to observe how the level of mindfulness of employees changes during the process of implementing and improving work gamification. Secondly, from the perspective of resources, work gamification is from the enterprise resource perspective, providing work guidance for employees, while workplace mindfulness is from the individual resource perspective, providing psychological guidance for employees. This is a parallel relationship, and it indeed reflects that the environment and individuals can influence behavior.(2.3)

To demonstrate that the correlation is not related to collinearity, this study conducted a multicollinearity diagnosis on each predictor variable. The results are shown that, with VIF ranging from 1.401 to 2.188, which is far less than 5; tolerance ranging from 0.457 to 0.714, which is much higher than 0.1. There is no serious multicollinearity problem in the data of this study.(4.3)

4.5. supplementary experiment

To make the model construction more complete and accurate, this study employed Model 6 in the process to explore the chain mediation. The results indicated that the chain mediation path of "digital leadership → work gamification → workplace mindfulness → employee resilience" was not significant, with an effect value of 0.019 and a 95% confidence interval of [-0.001, 0.048]. Since the 95% confidence interval included 0, it was more reasonable to set it as a parallel mediation.(4.5)

Future research could adopt panel data for long-term tracking to obtain more accurate research data and deeply analyze whether there are potential mediating paths between work gamification and workplace mindfulness, as well as the relationship between employee resilience and individual and environmental factors.(5.4)]”

Comment 7：[Seventh, we should expand the discussion on the non-significant results of the mediating role of AI anxiety regulating work gamification between digital leadership and employees' resilience. Because artificial intelligence anxiety is presented as a central explanation mechanism, its lack of significant mediating role deserves more adequate treatment. In view of the theoretical importance given to this construct in the manuscript, this is not sufficient. The authors should consider alternative explanations more carefully, including measurement problems, the limited specificity of AI anxiety measurement, and potential paths that may affect the main effect.]

Response 7：Thank you for your suggestion to conduct a more in-depth discussion on the non-significant results. AI anxiety does not moderate the path mediated by work gamification. The original text mentioned that digital leadership can effectively enhance employees' resilience levels through work gamification, and this path itself is quite stable. In addition, we also analyzed the reasons why AI anxiety does not moderate this mediating path from the limited heterogeneity of AI anxiety measurement and possible problems in the measurement. We believe that the scale of AI anxiety involves AI learning anxiety and job replacement anxiety, and the scale may not fully cover the content related to the work environment, so the moderation is not obvious. Moreover, it may also be related to data collection. The changes in digital leadership and work gamification may take time, and the moderation of AI anxiety on this path also requires time.

You can see this change on section 5.1 page 18 lines 642-658 of the revised manuscript.

“[Secondly, it can be seen from the limited specificity of the AI anxiety measurement. In this study, the scale for AI anxiety is oriented towards the two dimensions of "AI learning anxiety" and "job replacement anxiety", which is from the perspective of employees' psychology. However, digital leadership is a manifestation of a leader's management ability, and work gamification is a work system that involves the task design of an enterprise. Its impact on employee resilience is a kind of interesting experience, which is quite different from the psychological states of job replacement anxiety and learning anxiety. The content of the AI anxiety scale does not involve the aspects related to gamification and leadership, so the path of AI anxiety regulating through work gamification as a mediating variable is not significant. Finally, from the perspective of the data collection method, the data collected in this study is cross-sectional data. The moderating effect of AI anxiety on the path mediated by work gamification has a lag. It may be that the erosion of positive experiences by AI anxiety is slow, which is a gradual process. It is also possible that although AI anxiety is high, under the digital leadership of enterprise managers, the improvement of the work gamification system in the enterprise takes time, and the improved gamified work environment may be more conducive to enhancing employee resilience. Therefore, this path should be explored at multiple time points.(5.1)]”

Comment 8：[Finally, there are several editing and formatting problems that need to be corrected. These problems include the table cross-page does not set the continuation table, the chart position, the remaining template language, the reference format is not uniform and so on. While these issues are relatively easy to resolve, they give the impression that manuscripts are not yet fully prepared for submission in the final journal format.]

Response 8：Thank you for your valuable suggestions. We will improve the editing style and format of the article.

We have had our articles polished by native speakers and have taken the Wiley language test. The certificate is as follows.

Reviewer 2 Report

Comments and Suggestions for Authors

This study examines the underlying mechanisms through which digital leadership influences employee resilience by constructing a moderated mediation model, in which work gamification and workplace mindfulness serve as mediating variables, and AI anxiety functions as a moderating variable. The topic demonstrates strong practical relevance, particularly in the context of accelerated digital transformation and the increasing prominence of employee mental health issues. Methodologically, the study adopts confirmatory factor analysis, hierarchical regression, and Bootstrap techniques, reflecting a generally rigorous analytical approach. Overall, the manuscript shows potential; however, several critical issues remain at both the theoretical and methodological levels, which require substantial revision prior to publication.

First, the theoretical foundation requires further strengthening. Although the authors claim to base the study on Social Cognitive Theory, its application appears relatively superficial. Social Cognitive Theory emphasizes the concept of reciprocal determinism among environment, cognition, and behavior, whereas the current model is presented as a unidirectional linear relationship, failing to capture the dynamic interactions among these elements. Work gamification (as an environmental factor) and workplace mindfulness (as a cognitive factor) are modeled as parallel mediators without accounting for their potential interaction, which is inconsistent with the core logic of Social Cognitive Theory. It is recommended that the authors either incorporate feedback loops to enhance theoretical alignment or explicitly clarify that this study represents only a partial application of Social Cognitive Theory.

Second, there is a lack of coherence in the dual mediation model. The study assumes that work gamification and workplace mindfulness independently mediate the relationship between digital leadership and employee resilience. However, prior research suggests that these variables may be intrinsically connected in practice. For instance, a gamified work environment may enhance employee engagement and attentional focus, which in turn could influence their level of mindfulness. By overlooking this potential pathway, the current model weakens its theoretical explanatory power. The authors are advised to provide stronger theoretical justification for the parallel mediation structure or consider alternative model specifications.

Third, the theoretical positioning of AI anxiety remains unclear. The study proposes that AI anxiety exerts both enhancing and weakening effects on different paths, framing it as a “double-edged sword,” yet lacks a solid theoretical basis for this claim. Social Cognitive Theory alone may not sufficiently explain such a complex mechanism. It is recommended that the authors incorporate Conservation of Resources theory or the Job Demands–Resources model to conceptualize AI anxiety as a resource threat or job stressor, thereby offering a more systematic explanation of its dual effects.

Furthermore, the theoretical model (Figure 1) does not accurately reflect the proposed hypotheses. Specifically, AI anxiety is hypothesized and empirically tested as a moderator of three structural paths (DL→ER, WG→ER, and WM→ER), yet these moderating effects are not clearly depicted in the figure. As a result, there is inconsistency between the figure, the hypotheses, and the analytical model.

From a methodological perspective, the study faces concerns regarding common method bias and causal inference. All variables are derived from self-reported data collected at a single time point from the same source. Although the authors employed Harman’s single-factor test and a common latent factor approach, these methods are widely considered insufficient in contemporary research for adequately addressing common method bias. It is recommended that the authors adopt additional statistical remedies (e.g., marker variable techniques) or provide a more thorough discussion of this limitation. Moreover, the use of cross-sectional data does not support causal claims; therefore, causal language such as “influence” or “promote” should be revised to more appropriate terms such as “associate” or “relate to.”

There are also issues in the interpretation of results that warrant further clarification. For example, the direct effect of workplace mindfulness on employee resilience is non-significant in the regression analysis, while its indirect effect is significant in the Bootstrap analysis. This inconsistency is not adequately explained and may indicate a suppression effect or a conditional mediation mechanism, which requires deeper theoretical elaboration. In addition, the correlation between digital leadership and employee resilience is notably high (r = 0.844), raising concerns about potential construct overlap or common method bias. The authors are encouraged to report discriminant validity indices (e.g., HTMT) and address this issue explicitly.

Author Response

REVISION NOTES

Response to Reviewer 2 Comments

Summary

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[First, the theoretical foundation requires further strengthening. Although the authors claim to base the study on Social Cognitive Theory, its application appears relatively superficial. Social Cognitive Theory emphasizes the concept of reciprocal determinism among environment, cognition, and behavior, whereas the current model is presented as a unidirectional linear relationship, failing to capture the dynamic interactions among these elements. Work gamification (as an environmental factor) and workplace mindfulness (as a cognitive factor) are modeled as parallel mediators without accounting for their potential interaction, which is inconsistent with the core logic of Social Cognitive Theory. It is recommended that the authors either incorporate feedback loops to enhance theoretical alignment or explicitly clarify that this study represents only a partial application of Social Cognitive Theory.]

Response 1：Thank you for your suggestions for improving the theoretical framework of our study. We did apply a part of the Social Cognitive Theory in this research, and we have revised the original text to emphasize the use of "environment → individual → behavior". To make the model more complete, we also introduced the Conservation of Resources Theory to explain the moderating effect of AI anxiety and the differences in its moderating effects on different paths, making the existence of the moderating variable more reasonable. Additionally, we inserted phrases like "based on the Conservation of Resources Theory" in the original text to enhance the reading experience. Finally, in section 5.4, we pointed out the limitation of this study, which is the insufficient application of the triadic interaction relationship in the Social Cognitive Theory. We hope that future research can explore the core variables of this study more deeply through the triadic interaction relationship to see if employee resilience can reverse the influence of work gamification and digital leadership, etc.

You can see these changes on section introduction page 3 lines 113-117, section 2.4 page 8 lines 357-362, and section 5.4 page 22 lines 816-828 of the revised manuscript.

From the perspective of Social Cognitive Theory and Conservation Of Resources Theory, it can be understood as both a cognitive capability and a psychological resource. When AI anxiety is low, employees may not feel threatened in terms of work resources and feel little need to mobilize their cognitive resources. However, when AI anxiety is high, employees are more likely to perceive environmental instability and, in response, start looking for available resources and activate their mindfulness capacity in order to strengthen their resilience. (2.4 An explanation of the resource conservation theory has been added on the basis of the original text.)

Comment 2：[Second, there is a lack of coherence in the dual mediation model. The study assumes that work gamification and workplace mindfulness independently mediate the relationship between digital leadership and employee resilience. However, prior research suggests that these variables may be intrinsically connected in practice. For instance, a gamified work environment may enhance employee engagement and attentional focus, which in turn could influence their level of mindfulness. By overlooking this potential pathway, the current model weakens its theoretical explanatory power. The authors are advised to provide stronger theoretical justification for the parallel mediation structure or consider alternative model specifications.]

Response 2：Thank you for your valuable comments on the model construction of this study. We have provided additional explanations for why we chose parallel mediation. We believe that first, cross-sectional data is difficult to effectively support the influence of "digital leadership → work gamification → workplace mindfulness → employee resilience", because we think the impact of work gamification on workplace mindfulness is a dynamic one, and the influence of work gamification on workplace mindfulness needs to be studied through longitudinal data to explore how employee mindfulness changes during the establishment and improvement of work gamification. Secondly, from the perspective of conservation of resources theory, work gamification is an external environment, and workplace mindfulness is a personal resource. We want to explore the impact of environmental resources and personal resources on employee resilience through parallel mediation. In addition, to make the model more reasonable, we conducted supplementary experiments and also performed multicollinearity tests. The results showed that there was no chain mediation. Work gamification may have a potential impact on workplace mindfulness, but in this study, it does not weaken the explanatory power of the model. However, it can be explored in future research on the relationship between work gamification and workplace mindfulness.

Variable	Tolerance	VIF
DL	0.714	1.401
WG	0.520	1.921
AIA	0.466	2.146
WM	0.457	2.188

You can find these changes on section 2.4 page 9 lines 403-414, section 4.3 page 13 lines 506-510, section 4.5 page 17 lines 603-609, and section 5.4 page 21 lines 794-796 of the revised manuscript.

“[This study incorporates both work gamification and workplace mindfulness simultaneously into the mediating path, without establishing a chain mediation through which digital leadership influences employee resilience via work gamification and workplace mindfulness. The main reasons are as follows: Firstly, the data collected in this study is cross-sectional, which cannot effectively capture the chain mediating relationship between gamification and mindfulness. Chain mediation requires a clear temporal sequence, but the questionnaire data is difficult to observe how the level of mindfulness of employees changes during the process of implementing and improving work gamification. Secondly, from the perspective of resources, work gamification is from the enterprise resource perspective, providing work guidance for employees, while workplace mindfulness is from the individual resource perspective, providing psychological guidance for employees. This is a parallel relationship, and it indeed reflects that the environment and individuals can influence behavior.（2.4）

4.5. supplementary experiment

“[To make the model construction more complete and accurate, this study employed Model 6 in the process to explore the chain mediation. The results indicated that the chain mediation path of "digital leadership → work gamification → workplace mindfulness → employee resilience" was not significant, with an effect value of 0.019 and a 95% confidence interval of [-0.001, 0.048]. Since the 95% confidence interval included 0, it was more reasonable to set it as a parallel mediation.（4.5）

In the future, longitudinal data can be collected for tracking, and attention test questions can be added to the questionnaire design, or different scoring methods can be adopted for different variables to enhance the validity of the data.（5.4）]”

Comment 3：[Third, the theoretical positioning of AI anxiety remains unclear. The study proposes that AI anxiety exerts both enhancing and weakening effects on different paths, framing it as a “double-edged sword,” yet lacks a solid theoretical basis for this claim. Social Cognitive Theory alone may not sufficiently explain such a complex mechanism. It is recommended that the authors incorporate Conservation of Resources theory or the Job Demands–Resources model to conceptualize AI anxiety as a resource threat or job stressor, thereby offering a more systematic explanation of its dual effects.]

Response 3：Thank you very much for your valuable suggestions. In this study, we have introduced the resource conservation theory to make the effect of AI regulation more clear and explicit. According to the resource conservation theory, people will experience anxiety when they lose resources and may also seek new resources to maintain their existing ones. The negative moderating effect of AI anxiety on the direct path from digital leadership to employee resilience can be explained by the threat of resource loss in the resource conservation theory. The positive moderating effect of AI anxiety on the mediating path from digital leadership to workplace mindfulness to employee resilience can be explained by the motivation for resource protection and acquisition in the resource conservation theory. That is, when employees feel that their work resources are about to be lost, they will explore and utilize new resources, such as external digital technology resources, to enhance their mindfulness and use their psychological resources to regulate their stress, so as to better engage in work.

You can see this change on section introduction pages 3-4, lines 132-141.

“[According to the Conservation Of Resources Theory, resources are the driving force of human behavior and psychological states. When resources are at risk of being lost, people will experience stress and tension. AI anxiety is the negative emotion that employees feel when they are worried about losing their jobs due to the impact of AI. On the one hand, employees with high levels of AI anxiety will feel threatened when facing the digital leadership promoted by the enterprise, thus being immersed in tension and finding it difficult to engage in the learning of digital technologies. On the other hand, when the level of AI anxiety is high, employees will mobilize new resources to avoid losing their own resources and follow the organization to mobilize mindfulness resources to enhance their resilience. The manifestation of employee resilience is the result of resource mobilization.（introduction）]”

Comment 4：[Furthermore, the theoretical model (Figure 1) does not accurately reflect the proposed hypotheses. Specifically, AI anxiety is hypothesized and empirically tested as a moderator of three structural paths (DL→ER, WG→ER, and WM→ER), yet these moderating effects are not clearly depicted in the figure. As a result, there is inconsistency between the figure, the hypotheses, and the analytical model.]

Response 4：We are very grateful to the reviewers for pointing out the omissions in the details of the model. We had missed drawing the moderating arrow of AI anxiety on WG (Work Gamification) → ER (Employee Resilience) in the diagram. Now, it has been completed. These modifications can be seen in Figure 1.

Comment 5：[From a methodological perspective, the study faces concerns regarding common method bias and causal inference. All variables are derived from self-reported data collected at a single time point from the same source. Although the authors employed Harman’s single-factor test and a common latent factor approach, these methods are widely considered insufficient in contemporary research for adequately addressing common method bias. It is recommended that the authors adopt additional statistical remedies (e.g., marker variable techniques) or provide a more thorough discussion of this limitation. Moreover, the use of cross-sectional data does not support causal claims; therefore, causal language such as “influence” or “promote” should be revised to more appropriate terms such as “associate” or “relate to.”]

Response 5：Thank you for your valuable comments on methodological errors and writing. Although we distributed questionnaires in different quantities multiple times in our research, due to the cross-sectional nature of the data collected in this volume, errors are prone to occur. We acknowledged the limitations in 5.4 and also proposed improvement methods for the future. Additionally, we replaced causal-related terms in the abstract, and results with terms like "association" and "correlation" (changes are highlighted in blue).

You can see these changes on section 5.4 page 21 lines 777-781 and section 5.4 page 21 lines 792-796 of the revised manuscript.

“[First, to better investigate employee resilience, this study conducted questionnaire surveys on employees from different regions through two channels. It also carried out Harman's single-factor test and introduced a common method factor, but the common method bias was still difficult to eliminate（5.4）

Similarly, as this collection is cross-sectional data, it is impossible to completely rule out the potential impact of common method bias. In the future, longitudinal data can be collected for tracking, and attention test questions can be added to the questionnaire design, or different scoring methods can be adopted for different variables to enhance the validity of the data.（5.4）]”

Comment 6：[There are also issues in the interpretation of results that warrant further clarification. For example, the direct effect of workplace mindfulness on employee resilience is non-significant in the regression analysis, while its indirect effect is significant in the Bootstrap analysis. This inconsistency is not adequately explained and may indicate a suppression effect or a conditional mediation mechanism, which requires deeper theoretical elaboration. In addition, the correlation between digital leadership and employee resilience is notably high (r = 0.844), raising concerns about potential construct overlap or common method bias. The authors are encouraged to report discriminant validity indices (e.g., HTMT) and address this issue explicitly.]

Response 6：Thank you for your valuable comments on the explanation of the results and the correlation between variables. Regarding why mindfulness cannot directly affect employee resilience but can do so when it is used as a mediating variable in the workplace mindfulness path or under the moderating effect of AI anxiety, we believe that from the perspective of the conservation of resources theory, mindfulness can be regarded as a psychological resource, and employee resilience can be seen as a behavioral outcome of the continuous accumulation of resources. Psychological resources are difficult to directly influence employee resilience. From the perspective of resource protection, when employees sense a threat to their resources, they will mobilize resources, that is, use their own mindfulness ability to enhance resilience. From the perspective of resource acquisition in the conservation of resources theory, employees want to obtain more psychological resources, and at this time, the outside world can also provide ways for employees to acquire resources, such as digital leadership in this study, which can help employees improve their own mindfulness by introducing digital technology to the enterprise. At this point, employees will actively respond to the changes brought about by digital leadership to enhance their own mindfulness level, thereby improving their adaptability. Therefore, in this path, mindfulness is more like a resource converter. In addition, regarding the issue you raised about the high correlation between digital leadership and employee resilience, we conducted a multicollinearity test. With employee resilience as the dependent variable, the VIF result was far below 10, indicating no collinearity problem. Moreover, in the confirmatory factor analysis in the original text, we not only conducted a five-factor model test but also a four-factor model test. In Table 2, the four-factor model combined DL and ER, but the results showed that the four-factor model performed worse in RMSEA, CFI, and TLI, and CFI and TLI were even lower than the qualified value of 0.9. Therefore, we believe that digital leadership has no impact on employee resilience.

Thank you for your valuable comments. The specific changes can be seen on section 5.1 pages 18-19 lines 659-680 and on section 4.3 page 13 lines 506-510 of the revised manuscript.

Table 2. Confirmatory factor analysis results.

Model	χ²	df	χ²/df	RMSEA	CFI	TLI	SRMR
Five-factor: DL, ER, AIA, WG, WM	2212.968	1208	1.832	0.053	0.907	0.902	0.058
Four-factor: DL+ER, AIA, WG, WM	2342.602	1196	1.959	0.057	0.894	0.887	0.048

“[Furthermore, in this study, although workplace mindfulness cannot directly affect employee resilience, it can act as a mediating variable between digital leadership and employee resilience to exert an indirect influence on employee resilience. It will also be associated with employee resilience under the moderating effect of AI. A reasonable explanation is as follows: According to the conservation of resources theory, workplace mindfulness is a psychological resource, and employee resilience can be regarded as the behavioral outcome of resource accumulation or depletion, as well as an individual's adaptive result to the environment. According to the acquisition of resources in the conservation of resources theory, employees will mobilize their own resources for their own sense of security and also desire to continuously accumulate and obtain new psychological resources. Therefore, under the impact of AI anxiety, employees will constantly use their own mindfulness resources to change their resilience ability and also utilize external resources to change their mindfulness level. However, the mobilization of mindfulness level requires conditions, and it is difficult to increase or decrease by itself, so mindfulness is difficult to directly affect employee resilience. But when under the management of leaders with strong digital capabilities, enterprises may introduce digital technologies to enhance the mindfulness level of employees. To obtain more personal resources, employees will accept the digital resources within the enterprise and convert enterprise resources into their own psychological resources, thereby enhancing their mindfulness level, improving their adaptability, and increasing their resilience. Similarly, mindfulness is an uncritical acceptance of the present, and it does not have a clear action orientation by itself. However, employee resilience is a behavioral outcome, so mindfulness itself is difficult to have a significant association with employee resilience.（5.1）

We have had our articles polished by native speakers and have taken the Wiley language test. The certificate is as follows.

Reviewer 3 Report

Comments and Suggestions for Authors

The study itself is an interesting attempt to better understand underlying processes accounting for a relatively new set of dynamics at the tech oriented workplace - AI anxiety vs well-being through the concept of resilience. However a few aspects of the manuscript require improvement and revision before the manuscript may be published. These are specified herein:

At the opening paragraph of the intro - I would like to see some specific evidence to the extent of penetration of AI based tech into workplaces - try to show statistics: how many large employers, state-rum-offices etc. have officially embraced AI based tech into their workflow. How many of them are embracing a strategy of hiring people to train AI apps (which means that within a couple of years these "new hires" will not be needed anymore)? these should demonstrate the extent of the phenomenon that frames your study.
Frame your main goal early in the intro - if I got it right you basically are looking into the dynamics of anxiety/ stress vs. well-being in an AI saturated work setting. Say that. Respect the existing literature and discuss the notions associated with work related well-being before focusing on your specific variables.
In the literature review make sure you mention the leading authors in each field or concept e.g. digital leadership. Sources like Kane et al (2019); Contreras et al., (2020); Bolden & Regan and others do not get the place they deserve as the "founding fathers" of this concept, and so on.
The main issue I have with the manuscript is the theoretical model guiding it. Your outcome measure or variable is resilience. Resilience is the pattern of reaction that allows individuals and groups to cope effectively with challenges and return to normal function (and even surpass it) quickly. This means that resilience need to account for anxiety and stress, not the other way around. The authors must dedicate thought and rationale for the way variables are located in their model. A model that is more theoretically sound may suggest a double mediation in which : digital leadership leads to gamification and mindfulness (you will need to make a better case for how this happens), thus leading to higher employee resilience which in turn associates with AI anxiety levels. There are of course quite a few other options but the authors MUST create a clear rationale based on the theoretical and empirical evidence about the nature of each of their model's components.
More details are needed re the recruitment of participants - how were they recruited? from where?
In the results - under each table present a note with the full names of the variables to clarify the abbreviated names presented in the tables.
Using regressions to test for complex moderated-mediation model seems cumbersome . Why not use SEM or at least path analysis? Numerus software packages allow these analyses to be run in a seamless manner, which means less capitalization on chance and a better clarity of results.
This comment of course may become even more relevant should the authors accept my comment regarding the change in the study model.
In the discussion - avoid using causal language in the interpretation of the results (influence, cause, etc) since the study design is basically correlational - it is more fitting to discuss associations between variables.
The theoretical discussion is a bit problematic in my humble opinion - the framing of the results by social cognition theory is shaky - no central concepts in this field were actually measured in this study. The authors report measures associated more with emotional responses and resilience is actually a central pillar of positive psychology, coping models, etc. This inconsistency needs re-alignment.
The authors basically try to explain emotional responses and resilience in the hight of a workplace revolution. Coping theory, positive psychology and adaptation as well as classical leadership models seem naturally relevant here. Please revise this section.

Author Response

REVISION NOTES

Response to Reviewer 3 Comments

Summary

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[At the opening paragraph of the intro - I would like to see some specific evidence to the extent of penetration of AI based tech into workplaces - try to show statistics: how many large employers, state-rum-offices etc. have officially embraced AI based tech into their workflow. How many of them are embracing a strategy of hiring people to train AI apps (which means that within a couple of years these "new hires" will not be needed anymore)? these should demonstrate the extent of the phenomenon that frames your study. ]

Response 1：Thank you for your valuable suggestions. To demonstrate the extent of AI technology's penetration in the workplace and its impact on employees, and to present the research background, we have introduced the data from McKinsey's 2025 report, "The State of AI in 2025".

You can observe this change in section introduction ,pages 1-2 ,lines 40-44 of the revised manuscript and section introduction page 2 lines 45-48 of the revised manuscript.

“[According to the global survey report "The State of AI in 2025" released by McKinsey in 2025, 88% of the surveyed enterprises have routinely adopted artificial intelligence technology in at least one business. 42% of the enterprises plan to upgrade their employees' AI skills within 2025.(introduction)

According to the survey data from McKinsey (2025), 32% of the respondents predict that the total number of employees in enterprises will decrease by more than 3% due to AI. If this is the case, it can be seen that employees can feel the strong impact of AI technology.(introduction)]”

Comment 2：[Frame your main goal early in the intro - if I got it right you basically are looking into the dynamics of anxiety/ stress vs. well-being in an AI saturated work setting. Say that. Respect the existing literature and discuss the notions associated with work related well-being before focusing on your specific variables.]

Response 2：Thank you for your valuable suggestions. This paper sets AI anxiety as a moderating variable and, based on social cognitive theory and conservation of resources theory, constructs a model with digital leadership as the independent variable, work gamification and workplace mindfulness as the mediating variables, and employee resilience as the dependent variable. It can be said that this paper explores the relationship between digital leadership and employee resilience under the moderation of AI anxiety. Before discussing the specific variables in the literature review, we conducted a detailed analysis of the variable of employee resilience in the introduction, from the field to the selected scale and to the existing research on which variables are related to employee resilience. To highlight the field in which employee resilience is situated, we added a sentence in the original text.

You can see this change on section introduction page 2 lines 69-72 of the revised manuscript..

Comment 3：[In the literature review make sure you mention the leading authors in each field or concept e.g. digital leadership. Sources like Kane et al (2019); Contreras et al., (2020); Bolden & Regan and others do not get the place they deserve as the "founding fathers" of this concept, and so on.]

Response 3：Thank you for your valuable comments on the literature review. We have added the main authors for some variables in the original text. In the field of digital leadership, we have supplemented the literature of Kane et al. (2019) and Contreras et al. (2020) to better explain the evolution of the digital leadership variable in the early stage. We also cited the scale design of digital leadership proposed by Roma et al. (2019), which can help readers have a deeper understanding of digital leadership. In the field of work gamification, we have cited the literature of Deterding, S. et al. (2011) to introduce the concept of gamification and lay the foundation for the subsequent explanation of work gamification. We also integrated the original content to describe the common measurement methods under the Chinese scale. The literature citation in the field of workplace mindfulness is relatively complete, but we also briefly described the measurement methods of workplace mindfulness. In the field of AI anxiety, we have cited the literature of Li, J. and Huang, J. S. for explanation and presented the scale design and measurement methods of AI anxiety.

You can see this change on section 2.1 page 4 lines 163-171, section 2.1 page 4 lines 181-185, section 2.2 page 5 lines 209-212, section 2.2 page 5 lines 220-221, section 2.3 page 6 lines 275-278, section 2.4 page 7 lines 317-319, and section 2.4 page 8 lines 335-341

“[In the digital age, the changes and development of leadership have attracted considerable attention from scholars in recent years. Kane et al. were the first to study how digital disruption has transformed business and leadership, and put forward new requirements for traditional leadership skills (Kane G. C. et al., 2019). Contreras et al. proposed that in the special environment of the pandemic, an effective leadership is needed to guide employees in using electronic technology to enhance the adaptability of enterprises (Contreras Francoise et al., 2020). As digital technology gradually shifts from a technical support role to a key driver of organizational change, the concept and application of digital leadership have also been refined.(2.1)

the measurement of the degree of gamification within an enterprise’s internal work is usually determined by the number of gamification elements.(2.2)

Li, J., and Huang, J. S. hold that AI anxiety is the fear and concern of individuals about losing control due to the rapid development of AI technology(Li, J., &Huang, J. S.,2020).(2.4)

Kane, G. C., Phillips, A. N., Copulsky, J., & Andrus, G. (2019). How digital leadership is different. MIT Sloan Management Review, 60(3), 34-39.]”

Comment 4：[The main issue I have with the manuscript is the theoretical model guiding it. Your outcome measure or variable is resilience. Resilience is the pattern of reaction that allows individuals and groups to cope effectively with challenges and return to normal function (and even surpass it) quickly. This means that resilience need to account for anxiety and stress, not the other way around. The authors must dedicate thought and rationale for the way variables are located in their model. A model that is more theoretically sound may suggest a double mediation in which : digital leadership leads to gamification and mindfulness (you will need to make a better case for how this happens), thus leading to higher employee resilience which in turn associates with AI anxiety levels. There are of course quite a few other options but the authors MUST create a clear rationale based on the theoretical and empirical evidence about the nature of each of their model's components.]

Response 4：We are very grateful for your comments. In this article, we have provided a more comprehensive explanation of our model based on the "environment → individual → behavior" part of the social cognitive theory and the resource conservation theory. Firstly, according to the partial theory of the social cognitive theory, the digital leadership environment can influence the level of mindfulness and thereby affect employee resilience, and it can also change the resilience ability by altering the work gamification again. The reason for designing AI anxiety as a moderating variable in this study is that, according to the resource conservation theory, people will experience anxiety when they lose resources and may also seek new resources to maintain their existing ones. The negative moderation of AI anxiety on the direct path from digital leadership to employee resilience can be explained by the threat of resource loss in the resource conservation theory. The positive moderation of AI anxiety on the mediating path from digital leadership to workplace mindfulness to employee resilience can be explained by the motivation for resource protection and acquisition in the resource conservation theory, that is, when employees feel that their work resources are about to be lost, they will explore and use new resources, such as external digital technology resources, to enhance their mindfulness and utilize their psychological resources to regulate their stress, thereby better engaging in work. To make the positions of the entire model look more reasonable, we have also explained the mediating variables.

You can see these changes on section introduction page 3 lines 113-117, section introduction page 3 lines 117-121, and section introduction pages 3-4 lines 132-141 of the revised manuscript.

Comment 5：[More details are needed re the recruitment of participants - how were they recruited? from where?]

Response 5：Thank you for your suggestions on participant recruitment. We will improve this section by adding specific recruitment methods and locations.

You can see this change on section 3.1 page 10 lines 430-436 of the revised manuscript.

“[This study also recruited participants through two channels: Firstly, the questionnaire link was distributed to the internal staff of enterprises via the human resources department, allowing employees to fill it out themselves. Secondly, the questionnaire was distributed through online platforms, with the premise of the digital transformation background, to involve a wider range of workers. The distribution of the questionnaire was not limited to any specific region and covered employees from multiple provinces and cities.(3.1)]”

Comment 6：[In the results - under each table present a note with the full names of the variables to clarify the abbreviated names presented in the tables.]

Response 6：We are very grateful for your suggestions for improvement. Adding the full names of the abbreviations under each table makes the article clearer. We have made the additions in the tables.

You can see these changes on section 4 page 12 lines 484-485, section 4 page 13 lines 497-498, section 4 page 13 lines 513-515, section 4 page 15 lines 546 and 548, section 4 page 16 line 579, and section 4 page 17 line 601 of the revised manuscript.

“[Note: DL= Digital Leadership; WG = Work Gamification; WM = Workplace Mindfulness; ER = Employee Resilience; AIA = AI Anxiety.

Note: DL= Digital Leadership; WG = Work Gamification; WM = Workplace Mindfulness; ER = Employee Resilience; AIA = AI Anxiety.

Note:Digital Leadership; WG = Work Gamification; WM = Workplace Mindfulness; ER = Employee Resilience; AIA = AI Anxiety.

Note: DL= Digital Leadership; WG = Work Gamification; WM = Workplace Mindfulness.

Note: DL= Digital Leadership; WG = Work Gamification; WM = Workplace Mindfulness; ER = Employee Resilience.

Note: WG = Work Gamification; WM = Workplace Mindfulness; ER = Employee Resilience; AIA = AI Anxiety.

Note: DL= Digital Leadership; ER = Employee Resilience; AIA = AI Anxiety.（section 4）]”

Comment 7：[Using regressions to test for complex moderated-mediation model seems cumbersome . Why not use SEM or at least path analysis? Numerus software packages allow these analyses to be run in a seamless manner, which means less capitalization on chance and a better clarity of results.

This comment of course may become even more relevant should the authors accept my comment regarding the change in the study model.]

Response 7：Thank you for your better ideas on the research methods. In this article, we combined hierarchical regression with the Bootstrap method to test the model. The combination of these two methods is also a common approach in this field and can lead to the same conclusion as SEM. We will also further study the SEM method and strive to apply it to the next complex model test. Thank you again for your suggestions.

Comment 8：[In the discussion - avoid using causal language in the interpretation of the results (influence, cause, etc) since the study design is basically correlational - it is more fitting to discuss associations between variables. ]

Response 8：Thank you for your valuable suggestions. We have changed all the "influences" in the abstract and results to "correlations and exercises", and "positive influences" to "positive correlations".

Comment 9：[The theoretical discussion is a bit problematic in my humble opinion - the framing of the results by social cognition theory is shaky - no central concepts in this field were actually measured in this study. The authors report measures associated more with emotional responses and resilience is actually a central pillar of positive psychology, coping models, etc. This inconsistency needs re-alignment.

The authors basically try to explain emotional responses and resilience in the hight of a workplace revolution. Coping theory, positive psychology and adaptation as well as classical leadership models seem naturally relevant here. Please revise this section.]

Response 9：Thank you for your valuable suggestions. Workplace mindfulness and employee resilience are indeed core variables in positive psychology. The original text mainly studies the relationship between digital leadership and employee resilience. We want to start from the model of this study itself. First, because the variables in this study involve digital leadership, work gamification and other environmental factors, this is precisely linked to the "environment" in social cognitive theory. Workplace mindfulness in this study is a personal cognitive variable, thus linked to the "individual" in social cognitive theory. Finally, the meaning of employee resilience in this article is defined as an individual's adaptive behavior ability in the face of environmental changes, which is an employee's behavioral performance and linked to the "behavior" in social cognitive theory. Therefore, social cognitive theory can be used to construct the framework. Secondly, we regard AI anxiety as the manifestation of employees facing resource threats, mindfulness as personal psychological resources, and digital leadership as environmental resources, and explore the relationship between mindfulness and resilience from the perspective of resource conservation theory. Finally, positive psychology is indeed a good theory to explain mindfulness and employee resilience. We think that in the future, this theory can be used to explore what potential connections exist between mindfulness and employee resilience. Therefore, we introduced resource conservation theory in the article and acknowledged the use of some theories from social cognitive theory. We also used resource conservation theory to conduct an in-depth exploration of the relationship between mindfulness and resilience, and also introduced future research directions in 5.4.

Thank you again for your suggestions for improvement. You can find these changes on section introduction page 3 lines 113-117 and 132-141, section 5.1 pages 18-19 lines 659-680, and section 5.4 pages 21-22 lines 797-815 of the revised manuscript.

Furthermore, in this study, although workplace mindfulness cannot directly affect employee resilience, it can act as a mediating variable between digital leadership and employee resilience to exert an indirect influence on employee resilience. It will also be associated with employee resilience under the moderating effect of AI. A reasonable explanation is as follows: According to the conservation of resources theory, workplace mindfulness is a psychological resource, and employee resilience can be regarded as the behavioral outcome of resource accumulation or depletion, as well as an individual's adaptive result to the environment. According to the acquisition of resources in the conservation of resources theory, employees will mobilize their own resources for their own sense of security and also desire to continuously accumulate and obtain new psychological resources. Therefore, under the impact of AI anxiety, employees will constantly use their own mindfulness resources to change their resilience ability and also utilize external resources to change their mindfulness level. However, the mobilization of mindfulness level requires conditions, and it is difficult to increase or decrease by itself, so mindfulness is difficult to directly affect employee resilience. But when under the management of leaders with strong digital capabilities, enterprises may introduce digital technologies to enhance the mindfulness level of employees. To obtain more personal resources, employees will accept the digital resources within the enterprise and convert enterprise resources into their own psychological resources, thereby enhancing their mindfulness level, improving their adaptability, and increasing their resilience. Similarly, mindfulness is an uncritical acceptance of the present, and it does not have a clear action orientation by itself. However, employee resilience is a behavioral outcome, so mindfulness itself is difficult to have a significant association with employee resilience.（5.1）

Second, this study found that workplace mindfulness did not directly predict employee resilience; however, when workplace mindfulness served as a mediator between digital leadership and employee resilience, the mediation pathway was significant, and when the moderating effect of AI anxiety was taken into account, employee mindfulness is positively correlated with employee resilience.. The framework for the relationship between workplace mindfulness and employee resilience in this study is still not well established. Future research can further explore the conditions under which workplace mindfulness affects employee resilience. Moreover, as both mindfulness and employee resilience are core pillars in positive psychology, when studying the relationship between workplace mindfulness and employee resilience in the future, positive psychology can be introduced to construct relevant models, to more deeply explore the relationship between the two, and provide more effective micro-level suggestions on how to enhance employee resilience.（5.4）]”

We have had our articles polished by native speakers and have taken the Wiley language test. The certificate is as follows.

Reviewer 4 Report

Comments and Suggestions for Authors

The overall model is interesting, and the study has practical relevance for organizations undergoing digital transformation. The paper reviews relevant constructs, but the novelty of the study relative to existing work on digital leadership, technostress, employee resilience, and occupational health in digital settings is not yet sufficiently clear (and the introduction, and discussion, should state more explicitly what this study adds beyond combining known variables in one model)...

The methodological section should be strengthened: the authors should provide more detail on the sampling process, questionnaire administration, and steps taken to reduce potential bias: because all variables were collected through the same self-reported, cross-sectional survey, the risk of common method bias and causal overinterpretation remains important, especially given the very strong correlation between digital leadership and employee resilience.

Even the interpretation of results should be made more precise, where the paper should discuss more cautiously the finding that workplace mindfulness did not show a consistently significant direct effect on employee resilience in the regression model, even though the indirect effect was supported in the bootstrap analysis...

Comments on the Quality of English Language

There are many sentences with awkward phrasing, repetition, and stylistic imprecision; a careful revision by a proficient academic English editor would greatly improve readability and the presentation of the argument.

Author Response

Summary

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[The overall model is interesting, and the study has practical relevance for organizations undergoing digital transformation. The paper reviews relevant constructs, but the novelty of the study relative to existing work on digital leadership, technostress, employee resilience, and occupational health in digital settings is not yet sufficiently clear (and the introduction, and discussion, should state more explicitly what this study adds beyond combining known variables in one model)...]

Response 1：Thank you very much for your valuable comments. To highlight the novelty of this research, we have particularly emphasized the innovative points of this study from three aspects in the introduction: first, expanding the research level; second, combining social cognitive theory with resource conservation theory; and finally, exploring the moderating effect of AI anxiety. We also emphasized the contributions to social cognitive theory and resource conservation theory in 5.2 Theoretical Contributions and inserted resource conservation theory in the text.

These changes can be found on section introduction page 4 lines 145-152, section 5.2 page 19 lines 697-701, and section 5.2 page 20 lines 710-712 of the revised manuscript.

“[The innovation points of this research lie in the following aspects: First, it expands the research scope of digital leadership from the macro level of the organization to the micro level of employee resilience. Second, it combines social cognitive theory with resource conservation theory to construct a model and explains the differential moderating effect of AI anxiety. Third, it broadens the boundary conditions of employee resilience by incorporating AI anxiety as a moderating variable into the model, adapting to changes in the work environment and helping leaders enhance employee resilience through their own influence or internal organizational arrangements.（introduction）

Second, this study is based on the "environment → individual → behavior" path of Social Cognitive Theory and Conservation OfResources Theory. By integrating the structure of Social Cognitive Theory with Conservation Of Resources Theory to explain the dynamic changes of individual resources, it clarifies the process mechanism by which digital leadership influences employees' psychological resilience and the moderating role of AI anxiety.(5.2)

This study also analyzed the moderating effect of AI anxiety from the perspective of resource conservation theory and explored how to motivate employees to mobilize resources.(5.2)]”

Comment 2：[The methodological section should be strengthened: the authors should provide more detail on the sampling process, questionnaire administration, and steps taken to reduce potential bias: because all variables were collected through the same self-reported, cross-sectional survey, the risk of common method bias and causal overinterpretation remains important, especially given the very strong correlation between digital leadership and employee resilience.]

Response 2：Thank you very much for your comments on the methods section. We have refined the design of the questionnaire and the specific information on the distribution of the questionnaires (the way of distributing the questionnaires) during the research process. Regarding the very strong correlation between digital leadership and employee resilience, we conducted a multicollinearity test and found that the VIF was far less than 10, indicating no collinearity issue. However, we have also reflected on the shortcomings of this study and added them in 5.4, hoping to make improvements in future research.

Variable	Tolerance	VIF
DL	0.714	1.401
WG	0.520	1.921
AIA	0.466	2.146
WM	0.457	2.188

These changes can be found on section 3.1 page 10 lines 420-422, section 3.1 page 10 lines 430-436, section 4.3 page 13 lines 506-510, section 5.4 page 21 lines 777-780, and section 5.4 page 21 lines 792-796 of the revised manuscript.

“[The questionnaire emphasized that the data filled in this time would be used for academic research and would be in an anonymous form. There is no right or wrong answer for the filling results.（3.1）

This study also recruited participants through two channels: Firstly, the questionnaire link was distributed to the internal staff of enterprises via the human resources department, allowing employees to fill it out themselves. Secondly, the questionnaire was distributed through online platforms, with the premise of the digital transformation background, to involve a wider range of workers. The distribution of the questionnaire was not limited to any specific region and covered employees from multiple provinces and cities.（3.1）

……this study conducted questionnaire surveys on employees from different regions through two channels. It also carried out Harman's single-factor test and introduced a common method factor, but the common method bias was still difficult to eliminate.(5.4)

Comment 3：[Even the interpretation of results should be made more precise, where the paper should discuss more cautiously the finding that workplace mindfulness did not show a consistently significant direct effect on employee resilience in the regression model, even though the indirect effect was supported in the bootstrap analysis...]

Response 3：Thank you for your valuable comments on the explanation of the results and the correlation between variables. Regarding why mindfulness cannot directly affect employee resilience but can do so in the path with workplace mindfulness as the mediating variable and in the moderating environment of AI anxiety, we believe that from the perspective of the conservation of resources theory, mindfulness can be regarded as a psychological resource, and employee resilience can be seen as a behavioral outcome of the continuous accumulation of resources. Psychological resources are difficult to directly influence employee resilience. From the perspective of resource protection, when employees face resource threats, they are willing to use their existing resources, such as mindfulness (a psychological resource), to change their resilience. From the perspective of resource acquisition in the conservation of resources theory, employees want to obtain more psychological resources, and at this time, the outside world can provide ways for employees to acquire resources, such as digital leadership in this study, which can help employees improve their mindfulness by introducing digital technology to the enterprise. At this point, employees will actively respond to the changes brought about by digital leadership to enhance their mindfulness level, thereby improving their adaptability. Therefore, in this path, mindfulness is more like a resource converter.

These changes can be found on section 5.1 pages 18-19 lines 659-680 of the revised manuscript.

3. Response to Comments on the Quality of English Language

[We have carefully proofread the entire manuscript to improve the quality of English. In addition, the manuscript has undergone thorough English language editing by a professional editing service officially recommended by the journal.]

We have had our articles polished by native speakers and have taken the Wiley language test. The certificate is as follows.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Editor,

Thank you for the opportunity to review the revised manuscript.The authors have satisfactorily addressed all the concerns raised in the previous round of review. The revisions have significantly improved the manuscript, particularly in further refining the theoretical framework. I have no further comments or suggestions for revision. The paper is now suitable for publication in its current form.

Thank you again for the opportunity to contribute to the review process.

Best regards.

Author Response

Comment 1：Thank you for the opportunity to review the revised manuscript.The authors have satisfactorily addressed all the concerns raised in the previous round of review. The revisions have significantly improved the manuscript, particularly in further refining the theoretical framework. I have no further comments or suggestions for revision. The paper is now suitable for publication in its current form

Response 1：Thank you again for your valuable comments during the review process. We have revised the article based on your suggestions, making the theoretical framework clearer and improving the quality of the article. We are very grateful for your professional guidance during the review process, which has made our research more meaningful.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have responded to the reviewers’ comments in a careful and thorough manner. The revisions are well-considered and effectively address the concerns raised in the previous round. In particular, the authors have strengthened the theoretical grounding by incorporating additional perspectives, clarified the model specification, and provided more transparent explanations of the empirical results. Overall, the manuscript has been substantially improved in terms of clarity, coherence, and rigor.

Author Response

Comment 1：The authors have responded to the reviewers’ comments in a careful and thorough manner. The revisions are well-considered and effectively address the concerns raised in the previous round. In particular, the authors have strengthened the theoretical grounding by incorporating additional perspectives, clarified the model specification, and provided more transparent explanations of the empirical results. Overall, the manuscript has been substantially improved in terms of clarity, coherence, and rigor.

Response 1：Thank you for your full recognition and detailed evaluation of our revision work. We are deeply honored that the reviewer has acknowledged our efforts in strengthening the theoretical framework, clarifying the model setting, and explaining the empirical results. Your positive feedback is a significant encouragement to our research work. Once again, we would like to express our gratitude for the valuable time and professional guidance you have provided during the review process. Your valuable suggestions have enabled our research article to make a qualitative leap.

Reviewer 4 Report

Comments and Suggestions for Authors

The theoretical contribution should be clarified more precisely, and the claims of novelty should be stated more cautiously. The methodological section also needs stronger explanation, especially regarding the sampling process, the use of self-report cross-sectional data, and the treatment of AI anxiety as a single factor. In addition, some findings—especially the role of workplace mindfulness and the nonsupported moderated mediation through work gamification (should be interpreted more carefully and discussed with greater rigor).

Comments on the Quality of English Language

English requires further polishing before publication: there are recurring grammatical errors, awkward or overly literal phrasing, inconsistent tense usage, article/preposition problems, and several sentences that are unnecessarily long or repetitive.

In some sections, these language issues reduce precision and make the theoretical argument and interpretation of findings less clear than they should be.

Author Response

Summary

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[ The theoretical contribution should be clarified more precisely, and the claims of novelty should be stated more cautiously. The methodological section also needs stronger explanation, especially regarding the sampling process, the use of self-report cross-sectional data, and the treatment of AI anxiety as a single factor. In addition, some findings—especially the role of workplace mindfulness and the nonsupported moderated mediation through work gamification (should be interpreted more carefully and discussed with greater rigor).]

Response 1：Thank you for your valuable suggestions.

Regarding the need for a clearer presentation of theoretical contributions, we have divided them into three levels: Firstly, from the perspective of digital leadership influence, we have expanded from a macro perspective to a micro one. Secondly, in terms of theoretical integration, we have combined some social cognitive theories with resource conservation theory to propose a new framework. Lastly, we have expanded the boundary conditions of the employee resilience influence mechanism by introducing the moderating variable of AI anxiety.

You can see these changes in section 5.2, pages 18-19, lines 1655-1656, 1681-1683, 1690-1692 and 1700-1702 of the revised manuscript.

“[Firstly, this study expands the research perspective of digital leadership from the macro level to the micro level.(5.2)

Secondly, this paper combines social cognitive theory with COR theory, integrating individual and environmental factors into a unified framework, and expands the multiple paths through which digital leadership influences employee resilience.

This study also analyzed the reasons for the existence of the moderated mediation path and the direct path from the perspective of COR theory.(5.2)

Thirdly, this paper, based on employees' cognitive factors, incorporates AI anxiety into the research framework, thereby expanding the boundary conditions of the mechanism through which employee resilience exerts its influence.(5.2)]”

In the statement of the innovation points, we should be more cautious. We have changed "providing theoretical basis" to "research support", replaced words like "laying the foundation" and "solved", and substituted "opened up a new direction" with "provided a referential analytical framework".

These changes can be seen in section 5.2, pages 19, lines 1697-1699, and section 5.2, page 19, lines 1715-1717 of the revised manuscript.

“[……further strengthens the theoretical link between organizational support and employee health, and promotes interdisciplinary dialogue in the fields of occupational health and public health.(5.2)

provides a referable analytical framework for future research on the relationship between technostress and psychological resilience.(5.2)]”

In response to your suggestions regarding the methodology, we have added statements about the cross-sectional design in 3.1. The limitations of the cross-sectional data are mentioned in the original text in 5.4. “ Similarly, as this collection is cross-sectional data, it is impossible to completely rule out the potential impact of common method bias. In the future, longitudinal data can be collected for tracking, attention test questions can be added to the questionnaire design, or different scoring methods can be adopted for different variables to increase the validity of the data.”

These changes can be seen in section 3.1, page 9, lines 1040-1041 of the revised manuscript.

“[This study adopts a cross-sectional design, and all variables are collected at the same time point through the employees filling out questionnaires.（3.1）]”

Regarding sampling, in Section 3.1, we proposed the questionnaire design.”The questionnaire emphasized that the data filled in this time would be used for academic research and would be anonymous. There is no right or wrong answer for the filling results. ”and distribute questionnaires through two channels in Section 3.1“In this study, participants were also recruited through two channels. First, the questionnaire link was distributed to the internal staff of enterprises via the human resources department, allowing employees to fill it out themselves. Second, the questionnaire was distributed through online platforms, with the premise of the digital transformation background, to involve a wider range of workers. The distribution of the questionnaire was not limited to any specific region and covered employees from multiple provinces and cities. ”，The valid number of questionnaires and the proportion of sample characteristics were reported in section 3.1 .“In total, 400 questionnaires were distributed, of which 337 were returned. After 44 invalid or low-quality questionnaires were excluded, a final sample of 293 valid responses was retained, yielding an effective response rate of 73.25%. Among the respondents, 53.6% were male and 46.4% were female. Most participants were between 30 and 45 years old. In terms of educational background, 48.8% held a bachelor’s degree. Most respondents worked in enterprises with 51 to 500 employees. The sample also covers a wide range of enterprise ownership types, and 42.3% of the respondents were from firms located in eastern China.The detailed sample characteristics are presented in Table 1. ”，We have moved the statement "Detailed sample characteristics are shown in Table 1" to the end of the paragraph to align the textual content with the table. It is evident that the sampling process of this study is highly standardized and objective.

Regarding the treatment of AI anxiety as a single factor, we conducted an exploratory factor analysis, the specific content of which is presented in Section 3.2 of the original text.“The results of the exploratory factor analysis indicated that the six items of AI anxiety extracted one common factor (KMO = 0.877, cumulative variance explained rate = 63.974%), with item loadings ranging from 0.687 to 0.866, suggesting that it is feasible to treat AI anxiety as a whole.”，According to the research results of Li Jiang et al., there is no contradiction in the direction of the moderating effect. It may be due to the different scales selected that two dimensions in this study can be combined. If the dimension division is changed to different scales, new scales need to be selected for data analysis. The future research direction of AI substitution anxiety should be the same as that of Li Jiang et al., to refine the dimensions of AI anxiety and select different scales for more rigorous analysis. Therefore, in the original text, there is content related to EFA, and the possible shortcomings of this study can also be found in Section 5.4 of the original text.“Finally, the measurement of AI anxiety in this study has several limitations. This study selected AI anxiety, which includes AI learning anxiety and job replacement anxiety. However, subtle differences may exist between them. Future research should select two different scales for AI learning anxiety and job replacement anxiety to analyse the relationships between different dimensions of AI anxiety and different variables from a more microscopic perspective, making the research results more accurate.” The limitations are also mentioned in 3.2. Please refer to the main text and the limitations for guidance.

Li Jiang, Xuli Wang, Yuguang Xie, Junhong Zhu & Dongxiao Gu.(2026).Substitution or augmentation? How organisational AI adoption approaches influence the willingness to stay of caregivers.Computers in Human Behavior,174,108801-108801.https://doi.org/10.1016/J.CHB.2025.108801.

These changes can be seen in section 3.2, page 11, lines 1117-1118 of the revised manuscript.

“[Meanwhile, the limitations section discusses the theoretical differences between the two dimensions.]”

Mindfulness in the workplace does not have a direct relationship with employee resilience on its own, but it can be associated with employee resilience under the moderating effect of AI anxiety. It can also be related to employee resilience in a path where workplace mindfulness serves as a mediating variable. Through the resource conservation theory, mindfulness can be regarded as a psychological resource, and employee resilience can be seen as the behavioral outcome of the continuous accumulation of resources. Psychological resources are difficult to directly influence an employee's resilience. From the perspective of resource protection, when employees face resource threats, they are willing to utilize their existing resources, such as mindfulness (a psychological resource), to change their resilience. Therefore, under the moderating effect of AI anxiety, workplace mindfulness and employee resilience will establish a relationship. From the perspective of resource acquisition in the resource conservation theory, employees hope to obtain more psychological resources, and at this time, the outside world can provide employees with ways to acquire resources, such as digital leadership in this study. By introducing digital technology into the enterprise, it can help employees improve their mindfulness. At this time, employees will actively respond to the changes brought by digital leadership, enhance their level of mindfulness, and thereby improve their adaptability. Therefore, in this path, mindfulness is more like a resource converter. For specific explanations, please refer to the original text 5.1.

“Furthermore, in this study, although workplace mindfulness does not directly affect employee resilience, it can act as a mediating variable between digital leadership and employee resilience, exerting an indirect influence on employee resilience. It is also associated with employee resilience under the moderating effect of AI. A reasonable explanation is as follows: According to the conservation of resources theory, workplace mindfulness is a psychological resource, and employee resilience can be regarded as the behavioral outcome of resource accumulation or depletion, as well as an adaptive result to the environment. According to the acquisition of resources in the conservation of resources theory, employees will mobilize their own resources for their own sense of security and desire to continuously accumulate and obtain new psychological resources. Therefore, under the impact of AI anxiety, employees constantly use their own mindfulness resources to change their resilience and utilize external resources to change their level of mindfulness. However, the mobilization of mindfulness levels requires conditions, and it is difficult for it to increase or decrease by itself; thus, it is difficult for mindfulness to directly affect employee resilience. However, when under the management of leaders with strong digital capabilities, enterprises may introduce digital technologies to enhance employees’ mindfulness. To obtain more personal resources, employees accept digital resources within the enterprise and convert enterprise resources into their own psychological resources, thereby enhancing their mindfulness, improving their adaptability, and increasing their resilience. Similarly, mindfulness is an uncritical acceptance of the present and does not have a clear action orientation. However, employee resilience is a behavioral outcome; therefore, it is difficult for mindfulness itself to be significantly associated with employee resilience.”

Regarding work gamification, AI anxiety cannot moderate the path mediated by work gamification. The original text mentioned that digital leadership can effectively enhance employees' resilience levels through work gamification, and this path itself is very stable. In addition, we also analyzed the reasons why AI anxiety does not moderate this mediating path from the limited heterogeneity of AI anxiety measurement and possible problems in the measurement. We believe that the scale of AI anxiety involves AI learning anxiety and job replacement anxiety, and the scale may not fully cover the content related to the work environment, so the moderation is not obvious. Moreover, it may also be related to data collection. The changes in digital leadership and work gamification may take time, and the moderation of AI anxiety on this path also requires time. For specific explanations, please refer to 5.1 of the original text.

“A plausible explanation is as follows: First, in terms of the degree of perfection of work gamification. When leaders possess strong digital leadership, they tend to introduce gamification elements that match the organization’s culture and the overall characteristics of its workforce. A well-designed gamified work system enables employees to rapidly integrate into a positive work atmosphere, maintaining a transcendent positive psychological state within a “levelling up” institutional environment. When employees maintain a positive and upwards-oriented work mentality, they develop a robust capacity for stress management that preserves their learning ability and enhances their occupational health resilience. The entire pathway—from technological environmental change through institutional reform to behavioral influence—is inherently stable. Consequently, this pathway is relatively unaffected by AI anxiety, rendering the moderating effect negligible. Second, the specificity of the AI anxiety measurement is limited. In this study, the scale for AI anxiety is oriented towardthe two dimensions of “AI learning anxiety” and “job replacement anxiety”, which are from the perspective of employees’ psychology. However, digital leadership is a manifestation of a leader's management ability, and work gamification is a work system that involves the task design of an enterprise. Its impact on employee resilience is a kind of interesting experience that is quite different from the psychological states of job replacement anxiety and learning anxiety. The content of the AI anxiety scale does not involve aspects related to gamification and leadership; thus, the path through which AI anxiety is regulated through work gamification as a mediating variable is not significant. Finally, from the perspective of the data collection method, the data collected in this study are cross-sectional data. The moderating effect of AI anxiety on the path mediated by work gamification has a lag. It may be that the erosion of positive experiences by AI anxiety is slow, which is a gradual process. It is also possible that although AI anxiety is high, under the digital leadership of enterprise managers, improving the work gamification system in the enterprise takes time, and an improved gamified work environment may be more conducive to enhancing employee resilience. Therefore, this path should be explored at multiple time points.”

3. Response to Comments on the Quality of English Language

Comment 1：[English requires further polishing before publication: there are recurring grammatical errors, awkward or overly literal phrasing, inconsistent tense usage, article/preposition problems, and several sentences that are unnecessarily long or repetitive.

In some sections, these language issues reduce precision and make the theoretical argument and interpretation of findings less clear than they should be.]

Response 1：We thank the reviewer for the detailed and constructive feedback on the language quality of our manuscript. We acknowledge the specific issues raised—including grammatical errors, awkward phrasing, inconsistent tense usage, and article/preposition problems—and recognize that further polishing is required. The revised manuscript will undergo an additional round of professional English editing, with particular attention to the concerns highlighted in the reviewer's comments. We appreciate the reviewer's patience and guidance in helping us improve the clarity and precision of our writing.

We have had our articles polished by native speakers and have taken the Wiley language test. The certificate is as follows.

Round 3

Reviewer 4 Report

Comments and Suggestions for Authors

The literature review is somewhat repetitive and the theoretical contribution should be clarified more sharply, and the discussion should be more cautious, since the study is based on cross-sectional self-reported data and therefore does not support strong causal claims.

The interpretation of the findings also needs refinement: the role of workplace mindfulness should be explained more carefully, and the unsupported moderated mediation through work gamification should be discussed more explicitly.

I even think that the paper requires substantial English editing to improve clarity, fluency, and academic precision.

Comments on the Quality of English Language

English requires substantial editing to improve clarity, fluency, and academic precision because several sentences are awkwardly constructed or overly long, and there are recurrent grammatical, stylistic, and formatting inconsistencies that reduce readability.

Author Response

Summary

Thank you very much for taking the time to review this manuscript. Please find our detailed responses below. In the revised files, the content-related modifications are highlighted in purple, while the revisions made to improve the accuracy of the English language are shown using the Track Changes mode.

Point-by-point response to Comments and Suggestions for Authors

Comment 1：[The literature review is somewhat repetitive and the theoretical contribution should be clarified more sharply, and the discussion should be more cautious, since the study is based on cross-sectional self-reported data and therefore does not support strong causal claims.]

Response 1：Thank you for your valuable suggestions. Regarding the issue of some repetition in the literature, we have read through the article once. Apart from the background information and multi-angle introduction of variables from previous studies, we found that although some introductions cited different literature, they were explaining the same perspective. Therefore, we deleted them. In the text, we removed the causes of AI anxiety. As for the theoretical contribution needing to be more explicit, we deleted the part about theoretical contribution. The theoretical contribution of this study is more prominently highlighted as follows: Firstly, the research perspective of digital leadership is placed in a more micro field. Secondly, the utilization of social cognitive theory and resource conservation theory, as well as the construction of the framework. Finally, the introduction of the moderating variables of AI anxiety. In response to the claim that cross-sectional data does not have a strong causal relationship, we read through it again and checked the use of words.

You can see the wording changes in the revised manuscript in section 5.2, page 19, lines 680-681; section introduction, page 3, line 117; and section 4.4, page 14, lines 520-521.

“[……shifts the analytical focus of digital leadership from organizational performance to employees' psychological resources .(5.2)

……positively predicts……(1.introduction)

……it was indirectly associated with employee resilience through digital leadership.(4.4)]”

Comment 2：[The interpretation of the findings also needs refinement: the role of workplace mindfulness should be explained more carefully, and the unsupported moderated mediation through work gamification should be discussed more explicitly.]

Response 2：Thank you very much for your valuable suggestions. We have explained the insignificant paths related to workplace mindfulness and work gamification in section 5.1 of the article.

This is the explanation of the insignificant paths in the section 5.1 about workplace mindfulness.

This is an explanation of the insignificant paths in the section on work gamification in 5.1.

However, the results in 4. results were not well explained. To provide a simple explanation for the insignificant parts of the research, in 4.4, we started from the resource conservation theory to explain that the external factors of digital leadership and the moderating factors of AI anxiety are very important for the connection between workplace mindfulness and employee resilience. We also explained from the stability of work gamification why AI anxiety cannot moderate the path mediated by work gamification. To avoid redundancy, we only provided a simple explanation. The specific reasons are discussed in detail in 5.1 of the text.

You can see these changes in the revised manuscript in section 4.4, page 14, lines 524-527; section 4.4, page 15, lines 558-560; and section 4.4, page 16, lines 579-583.

“[This study indicates that digital leadership functions as an external environmental factor. As discussed in Section 5.1, it helps employees accumulate and mobilize psychological resources through supportive behaviors, thereby fostering employee resilience. (4.4)

This finding aligns with the logic of COR theory: AI anxiety, as a perceived resource threat, prompts employees to mobilize existing resources to enhance their resilience. (4.4)

This result may stem from the stability of work gamification. As determined by digital leadership, the institutional design of work gamification requires time for feedback and implementation. As an environmental factor, the enjoyment derived from gamification in the short term is also relatively stable and thus less susceptible to the moderating influence of AI anxiety.(4.4)]”

Comment 3：[I even think that the paper requires substantial English editing to improve clarity, fluency, and academic precision.]

Response 3：We thank the reviewer for the suggestion regarding language quality. The manuscript has undergone systematic language polishing to address issues of awkward sentence structure, redundant wording, and grammatical inconsistencies, thereby improving its clarity, fluency, and academic precision.

Response to Comments on the Quality of English Language

Comment 1：[English requires substantial editing to improve clarity, fluency, and academic precision because several sentences are awkwardly constructed or overly long, and there are recurrent grammatical, stylistic, and formatting inconsistencies that reduce readability.]

Response 1：Thank you for this suggestion. The manuscript has been professionally proofread and polished.

We have carefully polished and revised the entire manuscript for English language issues, and made thorough improvements throughout the text. Please find attached the certificate from our professional language editing service for your reference.