Surveillance, Disinformation, and Legislative Measures in the 21st Century: AI, Social Media, and the Future of Democracies
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe fact that AI is used by authoritarian regimes to control and persecute their own population can be conclusively proven by the author. However, the question arises as to why AI should always be viewed negatively in this context. Facial recognition can, for example, help to locate missing persons or fugitive criminals more quickly in public spaces. The crucial point here is not the technology itself, but the control of it. The distinction between authoritarian regimes and democratic systems and their measures for controlling artificial intelligence should therefore be made clearer. In addition, social media is mentioned in the title of the article, but the role of the major networks and their responsibility to control content is insufficiently mentioned.
The categorisation of the topics in the individual chapters is also not always coherent. For example, Chapter 2 discusses the use of AI to monitor the population. The last paragraph, however, is dedicated to disinformation campaigns. Lines 118 to 126 would probably belong in Chapter 3.
Chapter 4 provides a compact explanation of how micro-targeting works. But why is the role of large social networks such as meta not explained in more detail? This would also provide a better transition to Chapter 5, which deals in part with the attempted regulation of networks.
The conclusion is not coherent in relation to the overall content. Because when the author writes: Ultimately, AI serves as a tool that can both violate and uphold democratic principles in the contemporary world, a characteristic it is likely to retain in the future. As a reader, do you have to ask yourself where in the article AI is described as beneficial to democracy?
Useful sources for this could be:
10.1353/jod.2023.a907697
https://doi.org/10.1007/978-3-319-90869-4_7
https://doi.org/10.1177/20563051231186
Scientific standard practice:
Lines 88-90: a source must be cited here as to where the figures come from.
Lines 92-94: A quotation is reproduced directly here without a clear (page number) reference to the source
Throughout the text, care must be taken to ensure that direct quotations that are identified as such (using inverted commas) require a reference to the specific place in the source. See also: https://res.mdpi.com/data/mdpi_references_chicago_guide-v7.pdf
When a specific page, section, equation, or other division of the work is cited, it follows the date, preceded by a comma.
(Piaget 1980, p. 74)
(LaFree 2010, pp. 413, 417–18)
(Johnson 1979, sct. 24)
(Fowler and Hoyle 1965, eq. 87)
(Garcà a 1987, vol. 2)
(Smith 2001, chap. 1)
(Watson 2008, l. 3)
(Watson 2008, ll. 88–90)
(Lawrence 1999, para. 3–5)
(USCCB 1986, #14)
Lines 200-202 represent an evaluation by the author. This is usually only done in the conclusion.
Lines 213-214: A source should be used to demonstrate how it was proven that Russia influenced the 2016 US election.
The references have not yet been finalised. See for example (Ahmad et al. 2022)
Author Response
Responses to Reviewer 1
Comment 1: The fact that AI is used by authoritarian regimes to control and persecute their own population can be conclusively proven by the author. However, the question arises as to why AI should always be viewed negatively in this context. Facial recognition can, for example, help to locate missing persons or fugitive criminals more quickly in public spaces. The crucial point here is not the technology itself, but the control of it. The distinction between authoritarian regimes and democratic systems and their measures for controlling artificial intelligence should therefore be made clearer. In addition, social media is mentioned in the title of the article, but the role of the major networks and their responsibility to control content is insufficiently mentioned.
Response 1: As can be seen in the revised version of the article between lines 83-99, it was emphasized that it is also possible for AI to be used for good. The differences in goals/principles between democracies and authoritarian regimes were explained.
Comment 2: In addition, social media is mentioned in the title of the article, but the role of the major networks and their responsibility to control content is insufficiently mentioned.
Response 2: In its current form, lines 319-370 and 577-592 address the legal obligations of major networks regarding content removal and their various performances in this regard.
Comment 3: The categorization of the topics in the individual chapters is also not always coherent. For example, Chapter 2 discusses the use of AI to monitor the population. The last paragraph, however, is dedicated to disinformation campaigns. Lines 118 to 126 would probably belong in Chapter 3.
Response 3: The lines in question has been moved to the next chapter as per your recommendation. The information in question is now in the first sentence of Chapter 4. Please note that, due to the addition of a “methodology” section based on the suggestion of our other reviewer, what was previously Chapter 3 is now Chapter 4 in the current version.
Comment 4: Chapter 4 provides a compact explanation of how micro-targeting works. But why is the role of large social networks such as meta not explained in more detail? This would also provide a better transition to Chapter 5, which deals in part with the attempted regulation of networks.
Response 4: Chapter 4 from the previous version of the article has become Chapter 5 in this version due to the reason mentioned above. The former Chapter 5 is now Chapter 6. In the introduction of the current Chapter 5, Meta’s use of microtargeting is addressed.
Comment 5: The conclusion is not coherent in relation to the overall content. Because when the author writes: Ultimately, AI serves as a tool that can both violate and uphold democratic principles in the contemporary world, a characteristic it is likely to retain in the future. As a reader, do you have to ask yourself where in the article AI is described as beneficial to democracy?
Response 5: As you mentioned twice in your recommendation that the positive aspects of AI were not sufficiently addressed, we have added Chapter 8 to the current version of our article to discuss these points. Additionally, lines 577-592 (in the “Conclusions”) have also been newly added to the paper.
Comment 6: Useful sources for this could be:
10.1353/jod.2023.a907697
https://doi.org/10.1007/978-3-319-90869-4_7
https://doi.org/10.1177/20563051231186
Response 6: The source https://doi.org/10.1007/978-3-319-90869-4_7 was used (see lines 156-162). Unfortunately, the other links could not be accessed. We were prevented by the error “DOI NOT FOUND”.
Comment 7: Throughout the text, care must be taken to ensure that direct quotations that are identified as such (using inverted commas) require a reference to the specific place in the source
Response 7: Throughout the article, page numbers have been added to very specific citations/information that require them. See: line 115, line 120, line 124, line 130, line 131, line 157, line 165, line 202, line 232, line 257, line 269, line 341, lines 349-350, line 448, and line 454.
Comment 8: Lines 200-202 represent an evaluation by the author. This is usually only done in the conclusion.
Response 8: We did as you suggested. See lines 573-576.
Comment 9: Lines 213-214: A source should be used to demonstrate how it was proven that Russia influenced the 2016 US election.
Response 9: The Senate Select Committee on Intelligence (2018) is cited in the new version. See lines 284-287 in the current version.
Comment 10: The references have not yet been finalized. See for example (Ahmad et al. 2022)
Comment 10: We have crosschecked the references.
Please also note that we have split the chapter titled “4. Identifying the target audience for disinformation through the use of AI: Microtargeting” from the previous version into two. The new chapters, in line with their content, have been titled “5. Identifying the target audience for disinformation through the use of AI: Microtargeting” and “6. Combating the Misuse: AI and Content Removal”, respectively.
Many thanks for your invaluable and constructive contribution.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe paper provides a comprehensive overview of AI's role in political propaganda, surveillance, and efforts to combat misinformation. However, I was particularly concerned by the lack of a clear methodology section. I strongly recommend that the authors include details on how they conducted their literature review and selected their sources. In my view, incorporating more empirical data or case studies would significantly strengthen the key arguments presented. In terms of analytical depth, I believe some areas would benefit from more critical examination,like a deeper analysis of the effectiveness of legislative measures and the impacts of AI on propaganda and disinformation. I encourage the authors to provide a more nuanced analysis of these complex issues. The paper requires thorough editing for grammar, sentence structure, and clarity.
Comments on the Quality of English LanguageThe quality of English language in this paper is adequate but in need of significant improvement. There are numerous grammatical errors, awkward phrasings, mixed use of past and present tenses and inconsistencies that detract from the overall readability and professionalism of the work.
Author Response
Responses to Reviewer 2
Comment 1: The paper provides a comprehensive overview of AI's role in political propaganda, surveillance, and efforts to combat misinformation. However, I was particularly concerned by the lack of a clear methodology section. I strongly recommend that the authors include details on how they conducted their literature review and selected their sources.
Response 1: In this version of our article, we have added a methodology section based on your recommendation. We explained how we selected our sources and what we paid attention to while reviewing them. See lines 21-42.
Comment 2: In my view, incorporating more empirical data or case studies would significantly strengthen the key arguments presented.
Response 2: In this version of our article, between lines 257-274, arguments have been strengthened by including empirical data on microtargeting, between lines 339-350 on YouTube and Facebook, and between lines 446-454 on NetzDG.
Comment 3: In terms of analytical depth, I believe some areas would benefit from more critical examination, like a deeper analysis of the effectiveness of legislative measures and the impacts of AI on propaganda and disinformation. I encourage the authors to provide a more nuanced analysis of these complex issues.
Response 3: In the current version of our article, between lines 455-471, legislation related to the removal of content from social media has been discussed, and between lines 504-514, legislation on the protection of personal data has been examined in terms of its benefits and shortcomings. A similar discussion is also present in the conclusion section, between lines 573-592.
Comment 4: The paper requires thorough editing for grammar, sentence structure, and clarity.
Response 4: We have utilized MPDI’s language editing services.
Please note that, as the other reviewer criticized us for not sufficiently mentioning the benefits of AI, we have added Chapter 8.
Please also note that we have split the chapter titled “4. Identifying the target audience for disinformation through the use of AI: Microtargeting” from the previous version into two. The new chapters, in line with their content, have been titled “5. Identifying the target audience for disinformation through the use of AI: Microtargeting” and “6. Combating the Misuse: AI and Content Removal”, respectively.
Many thanks for your invaluable and constructive contribution.
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsIt was worth revising the article. It is now much more balanced and provides a comprehensive overview of the topic.
