Towards Smart Public Administration: A TOE-Based Empirical Study of AI Chatbot Adoption in a Transitioning Government Context

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study’s strengths include filling a gap in the literature on developed nations by employing the TOE framework in a novel way in a transitioning context. However, the use of self-reported data and a single-country sample may restrict wider applicability. A more thorough explanation of the survey instrument’s validation process and biases would strengthen the discussion. Furthermore, while the policy recommendations are sound, their implementation in limited resource contexts warrants more thorough consideration. The study has several methodological limitations that impact validity and generalizability. The first concern is self-reported survey data. This is a problem because respondents are likely to overstate their support for AI, leading to an artificial positive bias. The second problem is the cross-sectional design, which makes it impossible to track changes in attitudes over time or evaluate the long-term effects of adoption. The third problem is the focus on Uzbekistan. The study lacks consideration of the unique institutional barriers—such as bureaucratic inertia or infrastructural lack—unique to other transitioning contexts. The sample lacks a description of whether the selection process was random or purposive, which adds to the worries of representativeness if a concentration of respondents was from more technologically inclined departments.

 Measuring key constructs, such as "ethical AI regulation," remains vague, lacking specific indicators (e.g., compliance with international standards). Besides that, the absence of qualitative data (e.g., interviews) deprives the study of nuanced explanations of reasons for resistance or challenges to leadership. Despite this being an exploratory study for which PLS-SEM is suitable, its predictive power can be increased if additional robustness checks are undertaken or if comparative case studies from similar economies are investigated. Resolving these limitations with mixed methods, longitudinal data, or clearer sampling protocols would increase the study's rigor and generalizability to other transitioning governments.
Overall, this study makes a strong contribution to knowledge about AI adoption in public administration, especially in Central Asia. Some minor revisions, such as an elaboration of methodological limitations and contextual challenges, will increase its rigor and applicability. Finally, the following papers cannot fail to enrich this article and make it publishable.

https://www.tandfonline.com/doi/abs/10.1080/09638199.2017.1334809

Mahdavi, G.H., and Daryaei, A.A., (2015),"Giddens' Strctureation Theory, Corporate governance and
audit marketing",LudusVitalis, Vol. XI No. 1, pp. 1-7.

Author Response

We sincerely thank the editor for the opportunity to resubmit our manuscript, and we are also grateful to the reviewers for their constructive and insightful comments. All comments have been carefully considered, and corresponding revisions have been incorporated into the revised manuscript. Below, we detail our responses and the changes made.

Comments:
The study’s strengths include filling a gap in the literature on developed nations by employing the TOE framework in a novel way in a transitioning context. However, the use of self-reported data and a single-country sample may restrict wider applicability. A more thorough explanation of the survey instrument’s validation process and biases would strengthen the discussion:
Furthermore, while the policy recommendations are sound, their implementation in limited resource contexts warrants more thorough consideration.

Responses:
We appreciate the reviewer’s insight regarding the practical challenges of implementing policy recommendations in resource-constrained environments. In response, we have expanded our discussion to emphasize a phased, context-sensitive implementation strategy that accounts for institutional and resource limitations.

The revision is as follows:

In section 5.3 (page 23, line 885):

“...While the proposed policy recommendations offer a strategic roadmap for AI chatbot adoption, their implementation must be tailored to resource-constrained transitioning countries contexts. In transitioning governments, persistent limitations in funding, infrastructure, and human capital often constrain the implementation of large-scale reforms. A phased approach such as deploying regulatory sandboxes, adapting audit protocols to institutional capacity, and modular civil servant training programs ensure that effective adoption efforts remain both feasible and sustainable. Aligning policy ambitions with existing practical capabilities is essential for fostering accountable, ethical, and practical AI integration in public administration....”

Comments:
The first concern is self-reported survey data. This is a problem because respondents are likely to overstate their support for AI, leading to an artificial positive bias.

The second problem is the cross-sectional design, which makes it impossible to track changes in attitudes over time or evaluate the long-term effects of adoption.

The third problem is the focus on Uzbekistan. The study lacks consideration of the unique institutional barriers—such as bureaucratic inertia or infrastructural lack—unique to other transitioning contexts.

Despite this being an exploratory study for which PLS-SEM is suitable, its predictive power can be increased if additional robustness checks are undertaken or if comparative case studies from similar economies are investigated. Resolving these limitations with mixed methods, longitudinal data, or clearer sampling protocols would increase the study's rigor and generalizability to other transitioning governments.

Responses:
We sincerely thank the reviewer for these insightful and constructive observations. In response, we have revised the limitations and discussion sections to explicitly acknowledge and address these concerns:

• Self-reported survey data: We now explicitly discuss the potential for response bias, such as social desirability or overstatement of AI support, and recommend the use of triangulation with qualitative data in future studies to improve validity.
• Cross-sectional design: We have acknowledged the inability of our design to capture temporal shifts in perception or long-term outcomes, and suggest that future research adopt longitudinal approaches to strengthen causal inference.
• Single-country focus: While our study offers context-rich insights from Uzbekistan, we recognize that transitioning states face diverse institutional constraints. We have revised the discussion to reflect on common barriers in similar contexts and highlight the need for comparative research.
• Methodological suggestions: In line with the reviewer’s recommendations, we suggest that future studies employ mixed-methods designs, apply robustness checks, and enhance sampling protocols to increase the generalizability and predictive power of findings

The revision is as follows:

In section 5.4 (page 24, line 895):

“...This study has several limitations that should be acknowledged. First, reliance on self-reported survey data may introduce response biases, such as social desirability or overestimation of support for AI adoption. Second, the cross-sectional design limits the ability to capture changes in perceptions over time or assess long-term impacts of adoption. Third, the study focuses entirely on the case of Uzbekistan—while this provides valuable insights into a transitioning government context, its specific bureaucratic structure, digital maturity, and governance culture may differ from those in other transitioning states, where institutional trust, infrastructure limitations, or resistance to change may present distinct challenges. To strengthen the robustness and generalizability of future research, scholars are encouraged to replicate and extend this model in diverse country contexts, incorporate mixed-methods or longitudinal designs, and examine additional contextual moderators such as digital literacy, institutional trust, and regulatory environments...”

In section 5.4 (page 24, line 907):

“...Comparative studies across transitioning and emerging economies would also be instrumental in refining theoretical insights and expanding policy relevance. Finally, while this study applies PLS-SEM in an exploratory context appropriate for early-stage theory development, future research could enhance predictive power and generalizability by employing longitudinal or mixed-methods approaches, conducting robustness checks, and incorporating comparative case studies from other transitioning nations. Clearer sampling protocols and triangulation with qualitative data would further improve validity and contextual depth...”

Comments:
The sample lacks a description of whether the selection process was random or purposive, which adds to the worries of representativeness if a concentration of respondents was from more technologically inclined departments.

Responses:
Thank you for this insightful comment. We acknowledge the importance of clarifying the sampling strategy to address potential concerns about representativeness. In response, we have now clearly stated that a purposive sampling method was employed. The sample targeted public sector organizations actively engaged in digital transformation and AI chatbot initiatives to ensure that participants had relevant expertise and experience aligned with the study’s objectives.

This revision is made in Section 3.1 (page 12, line 507):

“...We employed a purposive sampling strategy, selecting public sector organizations involved in ongoing or planned AI chatbot initiatives and digital service delivery reforms. This ensured that participants possessed relevant knowledge and experience related to the topic…”

Comments:
Measuring key constructs, such as "ethical AI regulation," remains vague, lacking specific indicators (e.g., compliance with international standards). 
Besides that, the absence of qualitative data (e.g., interviews) deprives the study of nuanced explanations of reasons for resistance or challenges to leadership.

Responses:
We thank the reviewer for this insightful comment. In response, we have refined the operationalization of the construct “ethical AI regulation” by incorporating explicit indicators such as (i) the existence of national AI strategies or guidelines, (ii) alignment with internationally recognized standards (e.g., OECD AI Principles and UNESCO Recommendations), and (iii) presence of internal oversight and accountability mechanisms. These additions are reflected in the updated survey instrument and measurement model, as detailed in Section 3.1, 5.4 and Appendix A.

Regarding the absence of qualitative data, we acknowledge its importance in capturing deeper insights into resistance to leadership and organizational readiness. However, the current study was intentionally designed as a large-scale quantitative survey to identify generalizable patterns and test relationships across constructs. We recognize the complementary value of interviews and propose qualitative follow-up studies as an important next step for future research

This revision in Appendix A (page 30, line 1197):

“...The construct of ethical AI regulation was measured through indicators reflecting (i) the existence of national AI strategies or guidelines, (ii) alignment with OECD AI Principles and UNESCO’s Recommendation on the Ethics of AI, and (iii) the presence of internal oversight and enforcement mechanisms within public organizations (see Appendix A for measurement items). Items was clarified titled “Ethical AI Regulation (EAR)” to the survey instrument:

- EAR1: My organization follows national guidelines and international AI ethics standards (e.g., OECD, UNESCO) to use of AI.

- EAR2: Ethical concerns about AI chatbot adoption are actively addressed by top management.

 This revision is detailed in Section 5.4 (page 24, line 914):

“…Given the study’s quantitative design, we acknowledge the absence of qualitative data limits our ability to capture deeper organizational nuances. Future studies could complement these findings through interviews or case studies to explore factors such as resistance to leadership or institutional barriers in greater depth…”

Comments:
Finally, the following papers cannot fail to enrich this article and make it publishable.
https://www.tandfonline.com /doi/abs/10.1080/09638199. 2017.1334809

Responses:
We agree with the reviewer that governance quality is a fundamental driver of innovation adoption in public service delivery. This is further supported by Setayesh and Daryaei (2017), who demonstrate that good governance and innovation are mutually reinforcing in producing favorable economic outcomes, such as increased stock market turnover. Their findings strengthen our theoretical proposition that strong institutional governance is not only conducive to economic performance but also facilitates innovation-driven reforms, including the integration of AI technologies in the public sector. In the context of transitioning countries like Uzbekistan, where institutional reforms are ongoing, this linkage becomes particularly salient.

This revision is detailed in Section 5.1 (page 21, line 743):

“...This aligns with (Setayesh & Daryaei, 2017) results, which highlight how good governance and innovation jointly contribute to improved economic outcomes, thereby underscoring the importance of governance capacity in enabling AI-driven public sector transformation in transitioning contexts like Uzbekistan…”While this study focuses on Uzbekistan, similar institutional challenges such as bureaucratic inertia, fragmented infrastructure, and inconsistent digitalization policy enforcement are prevalent in many transitioning governments. These common barriers highlight the need for comparative cross-country research to better understand how such institutional factors influence AI adoption in public administration…”

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a timely and empirically grounded study of AI chatbot adoption in public administration using an extended Technology–Organization–Environment (TOE) framework. The integration of strategic governance variables into the TOE framework is theoretically innovative and expands its applicability to public sector digitalization. Overall, I found that this paper is generally well-structured and methodologically sound. It has the potential to contribute to both the AI governance and digital transformation literature. However, a few areas should be improved.

• The Uzbekistan context is a major strength, yet the empirical analysis does not fully capitalize on this unique setting. The authors should deepen the contextual interpretation of results by discussing how Uzbekistan’s institutional characteristics (e.g., centralized bureaucracy, digital governance strategies) interact with key findings.
• The paper claims a “novel contribution” by extending the TOE framework. However, the theoretical novelty remains under-elaborated. The authors may consider adding a subsection “Theoretical Contributions” to more clearly articulate how this study advances existing TOE literature, especially within the domain of digital public sector transformation.
• Additionally, the claim that this is the “first large-scale empirical study (p. 5)” on AI chatbot adoption in public administration in a transitional government context may be too strong. The authors may consider softening such statements or supporting them with a brief systematic review of prior TOE and AI chatbot adoption studies in comparable contexts.

 

Author Response

We sincerely thank the editor for the opportunity to resubmit our manuscript, and we are also thankful to the reviewers for their constructive and insightful comments. All comments have been carefully considered, and corresponding revisions have been incorporated into the manuscript. Below, we detail our responses and the changes made.

Comments:
This manuscript presents a timely and empirically grounded study of AI chatbot adoption in public administration using an extended Technology–Organization–Environment (TOE) framework. The integration of strategic governance variables into the TOE framework is theoretically innovative and expands its applicability to public sector digitalization. Overall, I found that this paper is generally well-structured and methodologically sound. It has the potential to contribute to both the AI governance and digital transformation literature. However, a few areas should be improved:
1. The Uzbekistan context is a major strength, yet the empirical analysis does not fully capitalize on this unique setting. The authors should deepen the contextual interpretation of results by discussing how Uzbekistan’s institutional characteristics (e.g., centralized bureaucracy, digital governance strategies) interact with key findings.

Responses:

We thank the reviewer for this insightful and constructive comment. In response, we have revised the Discussion section to deepen the contextual interpretation of our findings by linking them more explicitly to Uzbekistan’s institutional features. This revision aims to better reflect how the country’s centralized governance system and top-down digital transformation strategies shape the dynamics of AI chatbot adoption in public administration. Relevant policy initiatives have been referenced to support this contextualization.

The revision is as follows:

In section 5.1 (page 22, line 801):

“...Moreover, the institutional features of Uzbekistan particularly its transitional centralized bureaucratic structure and state-driven digital governance policy significantly mediate the adoption of AI chatbots. For instance, the emphasis on top-down digitalization strategies, such as the ‘Digital Uzbekistan – 2030’ initiative (Ministry for the Development of Information Technologies and Communications, 2020), may explain the relatively high influence of leadership and strategic alignment variables in our findings. Furthermore, the centralization of public administration could reinforce the salience of organizational readiness and accountability structures, as local agencies often rely on directives from the central government when adopting new technologies. These contextual factors highlight the importance of tailoring the TOE framework to transitional governance environments like Uzbekistan...”

Comments:
The paper claims a “novel contribution” by extending the TOE framework. However, the theoretical novelty remains under-elaborated. The authors may consider adding a subsection “Theoretical Contributions” to more clearly articulate how this study advances existing TOE literature, especially within the domain of digital public sector transformation

Responses:

e thank the reviewer for this valuable comment. We agree that the theoretical contribution needed to be articulated more explicitly. In response, we have added a dedicated subsection titled “Theoretical Contributions” after the Discussion section to clearly highlight how our study extends the TOE framework within the context of digital public sector transformation.

 

The revision is as follows:

In section 5.2 (page 22, line 832):

“...This study contributes to the theoretical advancement of the TOE framework by incorporating modern public governance-specific constructs such as leadership commitment, strong accountability, mindset change, strategic alignment, and ethical AI regulation into the ‘organizational’ and ‘environmental’ dimensions. Unlike prior applications of the TOE model that predominantly focus on private sector or generic institutional contexts, this research tailors the framework to address the distinct features of transitional governments. It responds to recent calls for contextualizing technology adoption models in the public sector by accounting for administrative hierarchies, institutional inertia, and normative governance expectations. By doing so, it extends the TOE’s explanatory power and offers a more robust framework for analyzing AI integration in digitally transforming public administrations, especially in developing country contexts...”

Comments:
Additionally, the claim that this is the “first large-scale empirical study (p. 5)” on AI chatbot adoption in public administration in a transitional government context may be too strong. The authors may consider softening such statements or supporting them with a brief systematic review of prior TOE and AI chatbot adoption studies in comparable contexts.

Responses:

We appreciate this important suggestion and agree that our claim requires more nuance. We have revised the relevant sentence. This strengthens the grounding of our claim and situates our work more accurately within the broader scholarly conversation.

 

This revision is detailed in Literature review (page 4, line 187):

“...According to the available evidence, this study is among the first empirical investigations to examine AI chatbot adoption in public administration within a transitioning government context, specifically through an extended TOE framework. While digital innovation in the public sector has received growing attention (Mergel et al.; Wirtz et al., 2019), and a number of studies have addressed AI use in government (Zuiderwijk et al., 2021), large-scale empirical research focused on chatbot-specific adoption in transitional public administrations remains limited. Recent systematic reviews (Sun & Medaglia, 2019) suggest that most studies on AI adoption in the public sector are concentrated in Western democracies. Therefore, this study contributes to filling both a geographical and theoretical gap by offering new empirical insights from the understudied Central Asian region, especially in the context of Uzbekistan...”

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

1) Improve the literature review section (Main)
In administrative research, the development of the study model is a critical component. In this study, variables were grouped under the categories of technology, organization, and environment; however, this was done without clear theoretical grounding. It is essential to specify the theories that underpin the selection and organization of these variables. Theoretical frameworks should be explicitly mentioned, critically discussed, and their limitations acknowledged. Only then can a well-justified conceptual model be constructed. The absence of a solid theoretical foundation represents a significant weakness in the current study. Include all relevant theories, provide a brief explanation of each, and offer a clear justification for the development of the proposed model.

2) The recruitment procedure of the participants is not clear (explain it in detail).
 
3) Include a summary table of the model fit indices to support the robustness of the structural model.

4) Report the R² and f² values to indicate the explanatory power and effect size of the model constructs.

5) Enhance the discussion section by comparing your findings with those of previous studies. Clearly explain the reasons behind the acceptance or rejection of each hypothesis within the context of your study.

Author Response

We would like to thank the editor for the opportunity to submit a revised manuscript, as well as the reviewers for their valuable and constructive comments, which helped us improve the quality of this paper. All of the issues raised by the reviewers are addressed in the revised manuscript. The changes and additions made to address the comments are outlined below.

Comments:
The study’s strengths include filling a gap in the literature on developed nations by employing the TOE framework in a novel way in a transitioning context. However, the use of self-reported data and a single-country sample may restrict wider applicability. A more thorough explanation of the survey instrument’s validation process and biases would strengthen the discussion:
Improve the literature review section (Main)
In administrative research, the development of the study model is a critical component. In this study, variables were grouped under the categories of technology, organization, and environment; however, this was done without clear theoretical grounding. It is essential to specify the theories that underpin the selection and organization of these variables. Theoretical frameworks should be explicitly mentioned, critically discussed, and their limitations acknowledged. Only then can a well-justified conceptual model be constructed. The absence of a solid theoretical foundation represents a significant weakness in the current study. Include all relevant theories, provide a brief explanation of each, and offer a clear justification for the development of the proposed model.

Responses:

Thank you for this valuable and constructive comment. We acknowledge the importance of providing a clear and robust theoretical foundation for the study model. In response to your feedback, we have undertaken a comprehensive revision of the Literature Review and hypothesis development sections to explicitly specify, justify, and critically discuss the theoretical underpinnings of the variables grouped under the Technology, Organization, and Environment (TOE) framework.

Specifically:

1. We have clearly articulated the TOE framework (Tornatzky & Fleischer, 1990) as the overarching theoretical lens guiding the structure of the study. The relevance of TOE to organizational-level technology adoption research, particularly in public administration settings, has been explained in detail.
2. To further consolidate the theoretical grounding, we have integrated complementary perspectives from the Diffusion of Innovation (DOI) theory (Orr, 2003) and the Technology Acceptance Model (TAM) (Davis, 1989). These theories are discussed not as standalone models but as supporting frameworks that inform specific constructs:
   • DOI informs constructs such as Compatibility and Perceived Relative Advantage, aligning with organizational readiness and contextual fit.
   • TAM is referenced to support perceptual variables like Perceived Usefulness and System Complexity (as a proxy for perceived ease of use).
3. We have critically discussed the limitations of DOI and TAM, emphasizing that while they offer valuable micro-level behavioral insights, they were not applied as primary models in this study due to their individual-level focus. Instead, TOE’s macro-level perspective was deemed more suitable for the institutional and organizational dynamics of public sector AI adoption. Additionally, the hypothesis development section has been revised to explicitly connect each construct to its theoretical origin, ensuring theoretical coherence throughout the model development process.

The revision is as follows:

In Literature review section (page 5, line 199):

“...This study adopts the Technology–Organization–Environment (TOE) framework as the primary structural foundation to categorize determinants of AI chatbot adoption in public administration. TOE offers a comprehensive lens for analyzing technological innovation adoption at the organizational level by considering internal and external contextual factors. To reinforce the conceptual robustness of the model, complementary theoretical perspectives have been incorporated. Specifically, the Technology Acceptance Model (TAM) informs the inclusion of perceptual constructs such as Perceived Usefulness and System Complexity (representing perceived ease of use). Despite TAM’s individual-level focus, its insights remain relevant for capturing user-oriented perceptions influencing organizational adoption decisions. Additionally, the Diffusion of Innovation (DOI) theory supports the inclusion of constructs like Compatibility and Perceived Relative Advantage, which reflect innovation characteristics crucial for assessing organizational readiness and contextual alignment in transitioning governments. By integrating TOE with TAM and DOI, this study develops a triangulated theoretical foundation that bridges macro-level organizational factors with micro-level behavioral attributes, thereby providing a comprehensive and context-sensitive model for AI adoption in public sector institutions. The limitations of each theory are acknowledged, and their combined application is justified to mitigate the conceptual gaps inherent in relying solely on a single framework...”


Comments:
The recruitment procedure of the participants is not clear (explain it in detail).

Responses:

Thank you for your valuable comment. We have revised the manuscript to provide a clearer and more detailed explanation of the participant recruitment process.

This revision is detailed in Section 3.1 (page 12, line 518):

“...Participants were recruited using a purposive sampling strategy targeting public sector employees involved in digital transformation initiatives across various government agencies in Uzbekistan. Initial contact was made an online through official channels, including institutional emails and professional networks. Inclusion criteria required participants to have at least one year of experience in public administration and familiarity with digital service tools or AI-assisted systems. A total of 501 valid responses were collected from those who voluntarily agreed to participate after being informed of the study’s objectives and confidentiality protocols...”

Comments:
Include a summary table of the model fit indices to support the robustness of the structural model.

Responses:

Thank you for the valuable suggestion. In response, we have added a summary table presenting key model fit indices (SRMR, NFI, d_ULS, d_G, and chi-square) generated using SmartPLS. This addition enhances the evaluation of structural model robustness and improves interpretability for readers. Furthermore, we reported the R² value and qualitatively assessed the f² effect sizes of model constructs in Section 4.2 to demonstrate explanatory power, given limitations in obtaining exact f² values from the software.

In Section 4.2 (page 16, line 631):

“...The model fit was assessed using multiple global fit indices. The Standardized Root Mean Square Residual (SRMR) was 0.085, indicating an acceptable fit below the conventional 0.10 threshold. Model is theoretically acceptable despite minor fit limitations. The Normed Fit Index (NFI) value of 0.771, although slightly below the commonly referenced threshold of 0.80, still indicates a marginally acceptable fit — particularly given the complexity of the model and the exploratory nature of the study. Additionally, the d_ULS (1.506) and d_G (1.126) values fell within acceptable ranges, further supporting overall model adequacy. The chi-square statistic (4332.929) is provided for reference only, as it is not directly interpretable within the PLS-SEM framework due to its sensitivity to sample size and model complexity...”

Table 6. Summary of Model Fit Indicators

Fit Index	Value	Threshold	Interpretation
SRMR	0.085	< 0.10	Acceptable model fit
NFI	0.771	≥ 0.80	Slightly acceptable
d_ULS	1.506	0	Acceptable
d_G	1.126	0	Acceptable
Chi-square	4332.9	—	For reference only
Collinearity (VIF)	< 3.3 for all constructs	< 3.3 (Kock, 2015)	No multicollinearity

Note: The SRMR value (0.085) falls below the recommended threshold of 0.10, indicating an acceptable level of model fit. Model is theoretically acceptable despite minor fit limitations. NFI is slightly below 0.80, suggesting a marginal but tolerable fit. Additionally, the collinearity statistics confirm that multicollinearity is not a concern, with all VIF values < 3.3.

In Section 4.2 (page 16, line 645):

“…Multicollinearity was not a concern, as all full collinearity VIF values were below the threshold of 3.3. The R² value for the intention to adopt AI chatbots (IAAC) was 0.73, indicating a substantial level of explained variance. Although exact f² values could not be extracted due to software limitations, the path coefficients indicate large effect sizes for perceived usefulness and compatibility; medium effects for effective accountability, traditional leadership, resistance to mindset change, and data management concerns; and smaller contributions from ethical AI regulation and system complexity. Taken together, these global fit indices and structural parameters confirm that the model demonstrates a reasonable and theoretically acceptable fit for explanatory purposes in the context of this exploratory study…”

Comments:
Enhance the discussion section by comparing your findings with those of previous 

Responses:

Thank you for the valuable suggestion. We have revised the discussion section to include a comparison of our findings with those of prior studies, particularly those applying the TOE framework in public sector and digital government contexts. This addition helps position our results within the broader literature and highlights the unique insights from the Uzbekistan case.

 

This revision is made in Section 5.1 (page 21, line 766):

“...Our findings corroborate prior research employing the TOE framework in public sector digital innovation studies (e.g., Alateyah et al., 2013; Jais et al., 2024), particularly regarding the significance of technological readiness and organizational support. However, distinct from studies conducted in more developed or stable environments, our results emphasize the amplified role of perceived usefulness, compatibility, system complexity, organizational readiness, leadership commitment, strong accountability, ethical AI regulation, concern over data management and security and resistance to digital mindset change within a transitioning government context. These findings reinforce the context-dependent nature of AI adoption drivers and suggest the necessity of adaptive, context-aware policy frameworks for emerging digital governments...”

Comments:
Clearly explain the reasons behind the acceptance or rejection of each hypothesis within the context of your study.

Responses:

Thank you for your constructive suggestion. In response, we have revised the manuscript to explicitly explain the rationale for the acceptance or rejection of each hypothesis. The revised section clearly links the statistical significance of the findings with theoretical expectations and contextual realities of Uzbekistan's transitioning public administration. This provides a more comprehensive interpretation of the results.

 

This revision is detailed in Section 4.3 (page 17, line 698):

“...The acceptance or rejection of each hypothesis was guided by both the statistical significance of the tested relationships and their theoretical relevance to the transitioning public administration context of Uzbekistan. Accepted hypotheses reflect strong empirical support consistent with prior literature and contextual dynamics, such as the influence of perceived usefulness, compatibility, organizational readiness, and effective accountability. Rejected hypotheses, such as those related to certain environmental or motivational variables, may be attributed to contextual factors like system complexity, traditional top-down leadership, resistance to mindset change, insufficient ethical AI regulation and concern over data management and security. These findings highlight the nuanced role of the TOE and extended variables in shaping AI adoption in a transitioning government context...”

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Accept in present form

Reviewer 3 Report

Comments and Suggestions for Authors

The paper has improved significantly compared to its previous version.