Understanding Factors Influencing Generative AI Use Intention: A Bayesian Network-Based Probabilistic Structural Equation Model Approach

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The use of data from 803 participants ensures robust statistical power and reliability.

Suggestions:

Clearly articulate how the integration of anthropomorphism and animacy enhances the existing literature on GAI adoption.

Expand on the rationale for selecting the EQ algorithm over alternative Bayesian network methods. Include a brief justification for its superior performance.

Detail how the clustering process was conducted to form constructs and their implications for the model's validity.

Provide additional evidence or references to support the inclusion of perceived anthropomorphism and animacy as constructs in GAI acceptance research.

Provide more details on the accuracy, recall, and precision metrics in Table 2. Are these values derived from cross-validation? what is the target variable. how did the authors encoded the responses? is there any correlation between variables?

Explain why performance expectancy had a lower impact compared to social influence, effort expectancy, and perceived usefulness. 

the Materials and Methods section should detail and include all the methods used in research, documented by literature. eq, variable clustering and so on.

Update and expand the literature review to include more recent studies on GAI adoption and Bayesian modeling.

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly. I have prepared a revised version that reflects the suggestions by the reviewers. I further addressed the reviewers’ comments one by one; a precise discussion of every feedback point can be found in the attached response letter.

I hope that the new version of the manuscript has satisfied your concerns, was enhanced enough to contribute to academia and the related field, and will help more future readers understand the research.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

- While the study adopts a Bayesian Network-based Probabilistic Structural Equation Modeling (PSEM) approach, it fails to establish a significant advancement beyond existing methods. The results largely reiterate previously well-documented factors such as social influence, perceived usefulness, and effort expectancy, offering limited new insights into generative AI adoption.

-  The reliance on a single dataset obtained through an online survey platform (Prolific) introduces potential sampling bias, as participants might not adequately represent the broader population of generative AI users.

- Despite mentioning the need for cross-cultural analysis, the research focuses exclusively on a single demographic, limiting its applicability to diverse user groups.

- The manuscript includes redundant explanations of theoretical models such as TAM and UTAUT, which do not directly contribute to the novelty of the study.

Comments on the Quality of English Language

The quality of English for this paper could be improved.

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly. I have prepared a revised version that reflects the suggestions by the reviewers. I further addressed the reviewers’ comments one by one; a precise discussion of every feedback point can be found in the attached response letter.

I hope that the new version of the manuscript has satisfied your concerns, was enhanced enough to contribute to academia and the related field, and will help more future readers understand the research.

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study offers an examination that includes a questionnaire analysis of the multifaced factors influencing the adoption of Generative AI. The research employs Bayesian Network-based Probabilistic Structural Equation Modeling. 

Please consider implementing the following suggestions for improvements:

1) Please explain why you used the Bayesian Network-based Probabilistic Structural Equation Modelling and compare it to other alternatives: " Unlike conventional methods...". The best way would be to add a table to the research that summarizes the differences between different models and approaches.

2) Please explain how you select the participants in the questionnaire.

3) Please provide more detailed examples of how you performed data cleansing, as mentioned in section 2.1.

4) Please explain why you used the Likert scale for questionnaire responses in comparison to other alternatives.

5) Please explain more details and characteristics related to the machine learning techniques that you mentioned here: "Before conducting comprehensive PSEM analysis, this study compared the performance of the EQ algorithm to other common machine learning techniques, including logistic regression, decision tree, support vector machine, neural network, random forest, AdaBoost, and bagging. ". It would be best to use separate subsections for each technique, explain the differences, and reference them properly.

6) Please explain Figures 1 and 2 in more detail: what every circle and color represents.

7) Please move this reference to the References section: "This research used an online survey platform Prolific (https://www.prolific.co)".

8) Similarly, please move this reference to the Reference section and define the version of the software that you used for your research: "utilized PSEM with Bayesialab 11 software (https://www.bayesia.com) to capitalize on its"

9) Some overlapping with external sources were not properly referenced. Please check the attached similarity report from TurnItIn. 

10) The number of references is not sufficient for such research, please add more references related to alternatives to Bayesian Network-based Probabilistic Structural Equation Modelling.

11) Please sort the abbreviations at the end of the manuscript alphabetically.

Comments for author File: Comments.pdf

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly. I have prepared a revised version that reflects the suggestions by the reviewers. I further addressed the reviewers’ comments one by one; a precise discussion of every feedback point can be found in the attached response letter.

I hope that the new version of the manuscript has satisfied your concerns, was enhanced enough to contribute to academia and the related field, and will help more future readers understand the research.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author improved the manuscript.

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly. 

Reviewer 2 Report

Comments and Suggestions for Authors

This study investigates the factors influencing users’ intention to use generative 9 AI by employing a Bayesian network-based probabilistic structural equation model ap- 10 approach. I appreciate the authors for the revision, but I only have the following minor concerns about the current version. In Section 2.3, when discussing the adoption of generative AI models in different areas. It is important to mention the safety of leveraging GAI, which could be maliciously exploited by adversaries. However, when discussing privacy issues, the current version has no citations to support its demonstration. Therefore, below are several references about using GAI to launch attacks and defenses you'd better add and discuss: (1) F2Key: Dynamically Converting Your Face into a Private Key Based on COTS Headphones for Reliable Voice Interaction (MobiSys’24), (2) AFace: Range-flexible Anti-spoofing Face Authentication via Smartphone Acoustic Sensing (UbiComp'24).

Comments on the Quality of English Language

The quality of English in this paper should be improved.

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly. I have prepared a revised version that reflects the suggestions by the reviewers.

As suggested by the reviewer, a new paragraph addressing the potential risks that need to be considered when adopting generative AI has been added to Section 2.3. References such as F2Key, which converts facial features into private keys using COTS headphones (MobiSys’24), and AFace, a smartphone-based anti-spoofing face authentication method (UbiComp’24), have been cited (please see page 4, line 182).

I hope that the new version of the manuscript has satisfied your concerns, was enhanced enough to contribute to academia and the related field, and will help more future readers understand the research.

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors, thank you for accepting my suggestions for improvements. As far as I'm concerned, the manuscript can be published now.

Author Response

I thank the reviewer for the timely and detailed feedback on the manuscript, and I believe that the feedback was much beneficial for us to improve the research significantly.