Chatbots and Empowerment in Gender-Based Violence: Mixed Methods Analysis of Psychological and Legal Assistance

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Peer review: Chatbots and empowerment in gender-based violence: Mixed methods analysis of psychological and legal assistance

Summary:

Overall, this is an incredibly strong piece of literature that offers fantastically insightful findings that are timely to the current climate of AI. This is very well written, and the author(s) have clearly put a lot of effort, time and energy into this piece. As a quantitative researcher, I would have liked to see grater inclusion of quotes from the interviews to bring more authority to your argument. However, I can easily see this qualitative data being used for a second paper focusing on the interviews. If the Author(s) haven’t considered this yet, I highly recommend it. While it may look like I have a lot of comments they are very minor, and the Author(s) should be incredibly proud of this piece. Its is strong, insightful and highly valuable. Thank you for the honour of reading this.

Abstract

• This is a very strong and informative abstract, however you just need to mention that you are looking at this in a Latin American context, as this has not been mentioned. Specifically, I recommend you including comments around. Your sample focusing on Peruvian women over the age of 18.
• Depending on word count availability, you could also mention the use of qualitative interviews and a 25 question survey.

Introduction

• Re consider use of the term “alarming”?
• Would be good to see a citation for line 92-93.
• Likewise, I recommend a citation for line 100-102
• The point about algorithmic bias (line 109-112 is a really interesting point, it would be good to see this expanded on slightly. For example, have other studies mentioned this and the impact of it?
• The justification of your research and the paper is fantastic! The introduction is also very well written and clearly shows a deep understanding of the topic.

Literature Review

• Proof reading error/repetition. Please fix this wording (line156) “Recent research indicates that Kim and Kwon's (2024) study indicates”
• Minor note: close the gap between 39 and % on line 174
• On line 186, you mention “In Brazil and Colombia, research has revealed that the algorithms used in these systems reflect gender biases that could reinforce stereotypes or provide erroneous advice, compromising the safety of victims (Guevara-Gómez, 2023; Avella et al., 2023).” This is incredibly interesting and would be good to see you expand on this slightly. What did Guevara-Gómez and Avella find specifically in relation to gender bias?
• Great to see the breakdown of hypotheses!

Materials and Methods

• Would be good to see a little mention of how many researchers were involved in the qualitative and quantitative analysis procedures. Aside from that, this is a very strong method with a clear and easy to follow structure!

Results

• Loved the visual components of this section – very engaging!
• The sentence on line 364 “This finding confirms that the effectiveness of chatbots in empowering users depends on digital access and trust in the technology” is a very bold statement, and at this stage in the results I don’t feel as though you have presented enough evidence to claim this. I suggest deleting it from here and moving it to the discussion/conclusion.
• Is the title for table 2 in a different language? Please review: “Resumen de hipótesis, variables, pruebas y resultados.”
• Line 458 you say “table 4” but I think you mean table 2? Please review.
• Title 4.2. Quantitative Analysis needs to be in italics

Discussion

• Would have been good to see a stronger focus on qualitative outputs, specifically it would have been good to see you include quotes from interviews with participants to bring further authority to your argument. Otherwise, there may be future opportunities to create an additional paper that focuses specifically on the qualitative interview outputs from this study as it holds highly valuable research.

Conclusion

• No comments, this is a very strong conclusion!

Author Response

Comment 1: Summary

This is a very solid and informative summary; however, you only need to mention that you are analyzing it in a Latin American context, as this has not been mentioned. Specifically, I recommend you include comments on this. Your sample focuses on Peruvian women over the age of 18.
Depending on word availability, you could also mention the use of qualitative interviews and a 25-item survey.
Response: Thank you for your comment. We have corrected the summary by explicitly mentioning Latin America and specifying that the sample consisted of Peruvian women over the age of 18. We have also added a reference to the use of qualitative interviews and the 25-item survey administered to the participants.

Comments 2: Introduction

Reconsider the use of the term “alarming”?
It would be good to see a citation for lines 92-93.
Likewise, I recommend a citation for line 100-102
The point about algorithmic bias (line 109-112 is a really interesting point, it would be good to see this expanded on slightly. For example, have other studies mentioned this and the impact of it?
The justification of your research and the paper is fantastic! The introduction is also very well written and clearly shows a deep understanding of the topic.
Response: We have addressed the comments made. We replaced the term “alarming” with more objective wording, added the corresponding citations in lines 92–93 and 100–102, and expanded the section on algorithmic bias by incorporating studies that demonstrate its impact on decision-making, human rights, and women's safety in Latin America.

Comments 3: Literature Review

Proof reading error/repetition. Please fix this wording (line156) “Recent research indicates that Kim and Kwon's (2024) study indicates”
Minor note: close the gap between 39 and % online 174
On line 186, you mention “In Brazil and Colombia, research has revealed that the algorithms used in these systems reflect gender biases that could reinforce stereotypes or provide erroneous advice, compromising the safety of victims (Guevara-Gómez, 2023; Avella et al., 2023).” This is incredibly interesting and it would be good to see you expand on this slightly. What did Guevara-Gómez and Avella find specifically in relation to gender bias?
Response: We have addressed the comments made. We corrected the wording in line 156 by removing the repetition, adjusted the formatting in line 174 by closing the space in “39%,” and expanded line 186 to incorporate the findings of Guevara-Gómez (2023) and Avella et al. (2023), along with additional references detailing how gender biases in artificial intelligence reproduce stereotypes, generate inappropriate responses, and compromise the safety of women in Brazil and Colombia.

Comments 4: Materials and Methods

It would be good to see a brief mention of how many researchers were involved in the qualitative and quantitative analysis procedures. Aside from that, this is a very strong method with a clear and easy-to-follow structure!
Response: We have addressed the noted observation. Section 3.5 now includes explicit mention of the number of researchers involved in each procedure. Thus, two researchers conducted the qualitative analysis using Atlas.ti, and three researchers specializing in statistical methods developed the quantitative analysis in RStudio.

Comments 5: Results

I loved the visual components of this section: very interesting!
The sentence on line 364, "This finding confirms that the effectiveness of chatbots in empowering users depends on digital access and trust in technology," is a very bold claim, and at this stage of the results, I don't think I have presented enough evidence to support this claim. I suggest removing it from here and moving it to the discussion/conclusion.
Is the title of Is Table 2 in another language? Please review: “Summary of Hypotheses, Variables, Tests, and Results.”
On line 458, you say “Table 4,” but I think you mean Table 2. Please review it.
Heading 4.2. Quantitative Analysis should be italicized.
Response: We have addressed the observations noted in the Results section. The statement on line 364 has been removed and moved to the discussion, the title of Table 2 has been corrected in English, the reference on line 458 has been adjusted from “Table 4” to “Table 2,” and the title of section 4.2 “Quantitative Analysis” has been italicized.

Comments 6: Discussion

It would have been positive to see more focus on qualitative results; specifically, it would have been positive if you had included quotes from participant interviews to bolster your argument. Otherwise, there may be future opportunities to create an additional article that focus specifically on the results of the qualitative interviews in this study, as it contains very valuable research.
Response: Thank you for your comment. I have included verbatim quotes from the

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a relevant and innovative study on the use of chatbots to support women affected by gender-based violence. The methodology is sound, the topic is highly significant, and the results provide empirical evidence for improving the design of chatbots with a gender-sensitive perspective. In addition, the bibliographic references used are up to date.

Below are some suggestions for improving the manuscript:

1. Move the definition of "chatbots" from section 1.1. "Contextualising artificial intelligence, chatbots and gender-based violence" to section 1. "Introduction" to clarify the concept for readers who are not familiar with it from the outset.

2. Specify in the Introduction how digital empowerment influences psychological and legal support.

3. Provide further justification for the final sample size of 1,000 participants. Sections 3.1 and 3.2 should be completed. In addition, it would be advisable to provide a better justification for the sampling method.

4. Expand on the information about the validation of the instruments (e.g., pilot tests, item structure).

5. Add the caption to Figure 1, as it currently reads: "This is a figure. Schemes follow the same formatting".

6. Revise the Discussion section. In many cases, this section repeats results rather than offering a more critical or theoretical analysis. Repetitive sections should be reduced to focus on theoretical implications and policy relevance.

7. Explore the ethical implications of the study in greater depth.

Once these suggestions have been incorporated, the manuscript will constitute a valuable contribution to the field of Social Sciences.

Author Response

The manuscript presents a relevant and innovative study on the use of chatbots to support women affected by gender-based violence. The methodology is sound, the topic is highly relevant, and the results provide empirical evidence to improve the design of chatbots with a gender perspective. Furthermore, the bibliography used is up-to-date.

The following are some suggestions for improving the manuscript:

Move the definition of "chatbots" from section 1.1. "Contextualizing Artificial Intelligence, Chatbots, and Gender-Based Violence" to section 1. "Introduction" to clarify the concept for those unfamiliar with it from the outset.
Response: Dear reviewer, thank you for your valuable suggestion. I have moved the definition of chatbots to the Introduction to make the concept clear from the outset of the manuscript.

Specify in the Introduction how digital empowerment influences psychological and legal support.
Response: Dear reviewer, we have added a paragraph to the Introduction (lines 43-51) that explains how digital empowerment influences psychological and legal support. It is noted that chatbots offer a safe space that reduces the fear of re-victimization while facilitating immediate access to legal resources and reporting channels. This reinforces the idea that digital empowerment is key to understanding the impact of these technologies.

Please justify the final sample size of 1,000 participants in more detail. Sections 3.1 and 3.2 should be completed. Furthermore, it would be advisable to further justify the sampling method.
Response: Comments regarding the justification for the sample size and sampling method have been addressed. Lines 332-338 explain the rationale for using 1,000 participants, highlighting the reduction in sampling error, the validity of the multivariate analyses, and the theoretical saturation achieved. Likewise, lines 347-354 describe the methodological strategy, which combines stratified sampling and purposive criteria within each stratum, with proportional quotas for urban/rural areas and connectivity levels, ensuring diversity and relevance of the selection.

Expand information on instrument validation (e.g., pilot testing, item structure).
Response: The comment was addressed by incorporating a description of the instrument validation process in lines 358-369. An explanation of the pilot test with 35 participants was added, along with the adjustments made to the items, the final structure of the questionnaire in five thematic blocks, content validation by a panel of experts, and internal reliability using the McDonald's Omega coefficient (ω = 0.89).

Add the title to Figure 1, as it currently reads: "This is a figure. The diagrams follow the same format."
Response: The corresponding title has been added to Figure 1 (lines 446-447), as follows: “Figure 1. Integration of AI chatbots with legal, psychological, and institutional support services for women,” thus correcting the noted observation.

Review the Discussion section. In many cases, this section repeats results rather than offering a more critical or theoretical analysis. Repetitive sections should be reduced to focus on theoretical implications and policy relevance.
Response: Relevant lines in the “Discussion” section have been adjusted, reducing the repetition of results and strengthening the critical analysis. The theoretical, policy, and ethical implications of the findings are now highlighted more clearly, in line with the recommendation.

Explore the ethical implications of the study in greater depth.
Response: I have added a new paragraph to the Discussion section (lines 634-643), which specifically addresses the ethical implications of the study. It highlights aspects such as confidentiality, protection of sensitive data, non-discrimination, empathy in responses, and prevention of re-victimization, with the aim of strengthening women's confidence in the use of chatbots and ensuring that technological innovation contributes to the protection of fundamental rights.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you very much for addressing the suggestions made in the first round of review. The manuscript shows substantial improvements and I greatly appreciate the effort that has been made. 

Author Response

In 3.3. Instruments, I incorporated Table 1. Categories, subcategories, indicators, and semi-structured questions, with their methodological justification. I specify that a pilot study of 30 interviews was conducted to refine the guide and that the qualitative analysis reported is based on 1,000 semi-structured interviews. In 4.1. Qualitative Analysis, I added Table 2. Coverage, subcodes, and microquotes by category and Table 3. Narrative summary by category and types of psychological and legal support, which organize the evidence by category and explicitly distinguish between psychological and legal support. I removed the word cloud due to its limited contribution and incorporated microquotes and coverage metrics to strengthen the qualitative approach.