Democratizing Artificial Intelligence for Social Good: A Bibliometric–Systematic Review Through a Social Science Lens
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear researcher/s:
It is a useful and interesting paper for focusing future researches centred in all this thematic fields discovered around AI and social goods. It has also very important results and conclusions, like the ones about the collaborations of humanities and SSCC researches with those of technological fields, the necessity of IA literacy, between the most important you have found in your meta-analysis. Furthermore, it is perfect structured and the methodological designed is correct. The use of Citespace demonstrates the tendency of the digital humanities in using digital research tools. Congratulations.
I only have to say these considerations:
- Maybe the word Democratizing is not the word that reflects the essence of the article. Without it will be perfect. As you mention in page 2, at the middle of the introduction, where you dedicated a space for this word and mention a few authors, "social good and democratization are inherently linked". This article talks about many things around IA researches in the Social Science fields. If you look at table 8, in your results, you wouldn't find democratization or a similar word. In the cluster's name, there is no democratization. In the top 20 keywords appears other words, like social good, governance and participatory technologies, that's all. Democratization was a keyword of your screening process in the database mentioned but is not a significant result of your work is a collateral word of social good. Its use looks reiterative in the title.
Another considerations for future research are that if you are talking several times about social problems and ethics, you have to consider the tendency in the academic world about making visible scientific women works. For example: you could say from the N=66 papers, in how many was a woman the principal author, and also in the different groups of themes. I know how complicated is this with the names in lots of different languages, you have to search for pictures of the researches...but for the next time with more time or help you can consider this.
In this point, before finishing, I have a question, N=66 (all the papers of the research) are in the bibliography references or where? Sorry if I haven't seen it if you mentioned in the text.
In line 56 there is an extra space you have to delete.
I hope I will see this work as soon as possible published. Sure, I will cite it.
Sorry for my English, it is mine, it is not NLS.
Author Response
Reviewer 1
Comment 1:
Maybe the word Democratizing is not the word that reflects the essence of the article. Without it will be perfect. As you mention in page 2, at the middle of the introduction, where you dedicated a space for this word and mention a few authors, "social good and democratization are inherently linked". This article talks about many things around IA researches in the Social Science fields. If you look at table 8, in your results, you wouldn't find democratization or a similar word. In the cluster's name, there is no democratization. In the top 20 keywords appears other words, like social good, governance and participatory technologies, that's all. Democratization was a keyword of your screening process in the database mentioned but is not a significant result of your work is a collateral word of social good. Its use looks reiterative in the title.
Response 1:
We appreciate your suggestion and understand the rationale behind it. However, we have decided to retain the term "democratizing," as it encapsulates a key aspect of the article's overarching theme. Nonetheless, we have enhanced the clarity of its usage within our context. As emphasised in the introduction, the connection between social good and democratization provides a critical lens through which the findings are discussed. While the term may not explicitly appear in Table 8, its conceptual relevance is deeply embedded in the narrative. Removing the term might overlook this foundational aspect of the discussion. To give more rationale about using democratization term, we add more explanation in the third paragraph (line 48-53) in Introduction section of revised manuscript. In short, this study adopts an inclusive operational definition of “democratizing AI for social good,” which essentially means ensuring equitable access to, participation in, and benefits from AI technologies for social good (line 68-71).
Comment 2:
Another considerations for future research are that if you are talking several times about social problems and ethics, you have to consider the tendency in the academic world about making visible scientific women works. For example: you could say from the N=66 papers, in how many was a woman the principal author, and also in the different groups of themes. I know how complicated is this with the names in lots of different languages, you have to search for pictures of the researches...but for the next time with more time or help you can consider this.
Response 2:
Thank you for your suggestion regarding making visible the scientific contributions of women and analyzing how many principal authors were women across different thematic groups. We recognize the importance of this perspective and acknowledge the contributions of women in AI for social good research in the manuscript. Specifically, we highlight notable works by authors such as Virginia Eubanks and Leila Ouchchy, emphasizing their impact on advancing social good and ethical discussions in AI. Please refer to the Section 4.2.4. Developing AI Literacy, Accessibility, and Social Inclusivity part in the revised manuscript.
Comment 3:
In this point, before finishing, I have a question, N=66 (all the papers of the research) are in the
bibliography references or where? Sorry if I haven't seen it if you mentioned in the text.
Response 3:
Thank you for your question. The 66 papers included in the research are summarized and structured within Section 3.2. They are also listed in the bibliography, intermingled with other references. We want to provide an additional online dataset for the 66 papers in a separate Excel document. We would like to seek the editor’s advice on whether it is suitable to attach this as an open dataset with the article, or if the current inclusion of the 66 articles in the references is sufficiently informative.
Comment 4:
In line 56 there is an extra space you have to delete.
Response 4:
Thank you for the comment. The extra space has been deleted.
Reviewer 2 Report
Comments and Suggestions for AuthorsThank you for this interesting and informative bibliometric-systematic review on the democratizing AI for social good. The paper gives a very good overview of the literature on AI for social good and discusses the mutlitude of concepts of and approaches to AI for social good very well.
However, I have a few remarks and issues with the paper, that in my opinion would need to be addressed before publishing, by this improving the content of the paper:
1. The introduction would significantly benefit from a clearer elaboration of what AI for social good consists of. While this concept will be detailed of course in the later part of the paper, currently the introduction is very superficial. Particularly the first and second paragraph don't contain anything substantial. Furthermore, a definition of vulnerable populations (line 23) would also be helpful. How is considered vulnerable to what extend and in what contexts?
2. In the methods section, I am missing a section on the limitations of the bibliometric and systematic review. Particularly as you are both using social science journals and computer science/natural sciences journals, journals and articles from the former might significantly lack in the comparison with other articles and journals. In such instances, there is a significant risk to overlook important (often critical) research from this field (e.g. from the field of surveillance studies or critical algorithm studies). Please reflect on this in your methods section.
3. Some of the findings are in my view a bit questionable. I mainly fail to see how marketing and customer relations in any way of thinking contributes to social good, even less through the means of AI. In almost every instance, the use of AI in this area is for profit generation and customer surveillance (see research on consumer surveillance or critical marketing studies). Furthermore, your findings could benifit from a discussion of the conflicting findings. no-code AI as well as information & news are both presented as contributing to social good as well as inhibiting social good. Discussing these points together would certainly help the reader to put this better into relation.
Finally, some minor issues I have encountered:
- p2 line 89-96: repitition of word "understanding".
- p3 line 108: Artificial Intelligence has already been abbreviated. Please use AI.
- p6 line 187: Table number wrong (2 instead of 3)
- p9 Table 8: What does the keyword "Article" actually contribute? Shouldn't this have been excluded?
- p20 line 673: Looks like a word is missing
I think addressing these issues would improve this otherwise very interesting paper.
Author Response
Reviewer 2
Comment 1:
The introduction would significantly benefit from a clearer elaboration of what AI for social good consists of. While this concept will be detailed of course in the later part of the paper, currently the introduction is very superficial. Particularly the first and second paragraph don't contain anything substantial. Furthermore, a definition of vulnerable populations (line 23) would also be helpful. How is considered vulnerable to what extend and in what contexts?
Response 1:
Thank you for your constructive feedback. The first and second paragraphs have been consolidated to provide a clearer emphasis on the role of AI in promoting social good. Furthermore, the statement regarding the benefits for vulnerable populations has been reinforced at the conclusion of the first paragraph (line 36) to underscore the significance of AI in advancing social good initiatives. Kindly refer to the revised manuscript for the updated first paragraph.
Comment 2:
In the methods section, I am missing a section on the limitations of the bibliometric and systematic review. Particularly as you are both using social science journals and computer science/natural sciences journals, journals and articles from the former might significantly lack in the comparison with other articles and journals. In such instances, there is a significant risk to overlook important (often critical) research from this field (e.g. from the field of surveillance studies or critical algorithm studies). Please reflect on this in your methods section.
Response 2:
Thank you very much for your valuable suggestion. We have incorporated the limitation of the bibliometric approach as recommended. Please refer to the fifth paragraph of Section 2.1 in the revised manuscript for the additional explanation.
Comment 3:
Some of the findings are in my view a bit questionable. I mainly fail to see how marketing and customer relations in any way of thinking contributes to social good, even less through the means of AI. In almost every instance, the use of AI in this area is for profit generation and customer surveillance (see research on consumer surveillance or critical marketing studies). Furthermore, your findings could benefit from a discussion of the conflicting findings. no-code AI as well as information & news are both presented as contributing to social good as well as inhibiting social good. Discussing these points together would certainly help the reader to put this better into relation.
Response 3:
Thank you for your valuable comment. We have emphasized the connection between marketing and social good by highlighting its relevance to vulnerable populations and small and medium-sized enterprises. For more details, please refer to Section 3.2.5.2, Marketing and Customer Engagement in the revised manuscript. To address the worries about fake information and the side effects of no-code AI, we also add a paragraph about critical AI literacy in Section 4.2.4.
Comment 4:
Finally, some minor issues I have encountered:
- p2 line 89-96: repetition of word "understanding".
- p3 line 108: Artificial Intelligence has already been abbreviated. Please use AI.
- p6 line 187: Table number wrong (2 instead of 3)
- p9 Table 8: What does the keyword "Article" actually contribute? Shouldn't this have been
excluded?
- p20 line 673: Looks like a word is missing.
Response 4:
Thank you for your suggestion. We have paraphrased to address the repetition of the word "understanding" and adjusted the abbreviation of Artificial Intelligence for consistency. Regarding Table 8, we have reordered the keywords based on centrality and excluded the term "article." Additionally, the sentence previously on line 673 has been revised and now appears on line 684. Finally, typos are corrected, and language expressions are improved in the revised manuscript.
