Talk:Algorithmic amplification/GA1

GA review

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Article (edit | visual edit | history) · Article talk (edit | history) · Watch

Nominator: Metalicat (talk · contribs) 17:37, 11 March 2026 (UTC)

Reviewer: A.Cython (talk · contribs) 22:07, 24 March 2026 (UTC)

I will review this as part of GARC. Expect comments in the following days. A.Cython(talk) 22:07, 24 March 2026 (UTC)

It was an interesting article without a doubt and should be improved to get to GA status. and I believe Metalicat made a valiant effort to bring this article to the current level. However, I have some concerns about the article's compliance with WP:MOS. Nothing major to result in a quickfail, but there are several minor and moderate issues that demand attention.A.Cython(talk) 20:08, 26 March 2026 (UTC)

Various

• No edit war
• Neutral
• No copyright violations, Earwig's Copyvio Detector gives 27.5%
• Figures have appropriate justification (created by the nominator) and figure captions

MoS

• remove citations from the lead per MOS:LEAD - Done.

• remove , according to Jian Xu from the lead. The lead is a summary of the main body and it should free of opinions. - Done.
• add wikilink, once in the lead (if mentioned), and once in the main body (first instance)
  • content moderation - Done.
  • recommendation algorithm - Done.
  • e-commerce - Done.
  • AI - Done.
  • Pearson - paragraph removed (see Sources below).
  • Social bot - Done.
  • whistleblower - Done.
  • civil discourse (note that you wrote: civic discourse)

Not done, I think, "civic discourse" is correct here. The DSA and European Commission use "civic discourse" for this context. Let me know your thoughts?

• • child safety - Done.
  • mental health - Done.
  • clickbait - Done.
  • news outlets - Done
  • Sock-puppet - Done.
  • oversight - Done.
  • SAFE For Kids Act (not that you have: SAFE for Kids Act) - Done
• when you mention "platforms" should it not wikilinked to Social media, as platforms.

not done, the article deliberately covers platforms beyond social media (Spotify, Amazon, adult content platforms), so I think linking "platforms" to Social media would be misleading given the article's scope.

• should it be For You be in Italic

Done.

• move wikilink of machine-learning to its first instance, i.e., developed machine-learning systems

Done.

• cultural intermediary does not need italics

Done

• At "See also" remove any links that are mentioned in main body per MOS:ALSO As a general rule, the "See also" section should not repeat links that appear in the article's body

Done.

• Define the acronyms at the first instance, once defined then use them, i.e., by replacing the name, e.g., DSA.

Done.

Prose

• Algorithmic recommendation either change "recommendation" → "recommendations" or add "approaches"

Done.

• accounts in the dataset delete "accounts in the" a bit redundant as mentioned before "which"

Done, sentence restructured as part of the broader prose changes.

• The regulatory framework developed add "was" before "developed"

Done.

• algorithmic curation is often used is "often" needed, reduce such words whenever possible per WP:WEASEL

Done.

• described the system as one of the earliest please be more explicit which "system" is that, the one of Amazon?

Done clarified in article.

• Please reduce the instances of single sentence paragraphs by either expanding them or merging them with nearby paragraphs.

Done, merged throughout.

• Explain what is a "Social bot" the current description is too vague function as amplifiers... not sure what to make of it.

Done, now defined on first use as "automated accounts that mimic human behaviour on social media platforms".

• rewrite by avoiding subjective terms
  • most direct evidence remove "most" - Done.
  • very large content remove "very" - Done. Note that "very large online platforms" remains in the Regulation section as this is the official legal term used in the DSA itself.
  • highly dependent remove "highly" - Done.
• According to Xu you should clarify who is Xu since you mention several names with Xu, e.g., Xu Xu, Yiqing Xu, and Jian Xu, however, in this instance it should altogether removed, we avoid the use of opinions (there are exceptions, but these are exceptions not the rule). - Done. Clarified to "Jian Xu" in the China section where the attribution is retained; removed from the lead.

• Another issue that I have noticed is the following: You frequently say YYY scholar did ZZZ study. This is academic style of writing and it is against WP encyclopedic style, per WP:NOTGUIDE#6 & 7. Try to summarize the relevant topic (remove names of the authors as a casual reader would not know them, remove jargon, emphasize on the relevant concepts). Ideally, you want to use academic reviews not papers per WP:MEDRS Cite reviews, don't write them. this is one of the guidelines most often abused

Partially done - Fair point on the prose style. I slipped into my Uni days! I've restructured throughout so that findings lead the sentence and attribution follows or sits in the citation. I have retained author names only where the researcher is independently notable and wikilinked (Zeynep Tufekci, Shoshana Zuboff, Frances Haugen, Georgina Born) or where the attribution is editorially necessary (Jian Xu in the China section, Huszár et al. where the study is referenced by name elsewhere in the article). On WP:MEDRS, I'd note that this guideline applies to biomedical claims; the article's subject falls within media studies and computer science, where id contend that citing individual studies rather than reviews is standard practice. I've addressed the underlying prose concern by restructuring rather than by changing the sourcing model. Happy to discuss further.

Sources

Spot check: TODO later 8, 9, 11, 12, 13, 16, 36, 40, 60, 62

• The paragraph starting Recommender systems can also encode relies on a single preprint. This is not acceptable per WP policies until this is peer-reviewed or add a different citation to support its claims. Otherwise this is not considered WP:RS and raises issues of WP:PROMO since anyone can write anything and upload it on archive server. Once published then it can be re-introduced.

Agreed. I have removed all material sourced to the arXiv preprint. The Mechanisms paragraph, the privacy/profiling argument in Criticism, and the constrained recommendation point have all been cut. The Methods section now cites only the peer-reviewed 2024 Applied Network Science paper by the same author.

• [AC] I also recommend to remove two citations on Spotify (#11 & #12) per WP:PROMO, they read as promotional. One citation from the company is more than sufficient, though it would be preferable this to come from an independent source, e.g., Guardian or academic.
• [AC] The same goes with citation 16 about the history of Amazon's recommendation algorithm from an Amazon news website. This is a clear conflict of interest with regard to the reliability of the particular source. You have an additional reference from a different source to support this so this removal does not cause any damage.
• [AC] When you have a paragraph with two or three citations but all are the same, then there is no need to cite each sentence with the same citation, just at the end, see WP:CONSECUTIVECITE. I think I saw two such paragraphs.

Summary

Hello A.Cython and thanks for the thorough review. Im glad you enjoyed reading the article. I think I have addressed all the points so far. Metalicat (talk) 00:58, 27 March 2026 (UTC)

@Metalicat Apologies for being a bit late slow as I was distracted by obligations. About the review. Thank you for making changes and you provided a good reason when you disagreed; it is not unusual for me to be wrong. Overall, the article has been improved positively. Unfortunately as I was doing the Spot check of the sources, I have a few more comments that require your attention. You have done a great job assembling the sources, most look great, but not all. See under the "Sources" subsection about the details. Once you have addressed these concerns I will conclude the review. A.Cython(talk) 04:13, 31 March 2026 (UTC)

@A.Cython No worries at all and thanks very much for the thorough review. I believe I have addressed the refereeing issues you have raised. Also thank you for the source suggestions in your last post. Metalicat (talk) 17:32, 31 March 2026 (UTC)

Final comments

Thank you for making the changes. I think there is a bit more room for improvement (better summaries), but given that this is an emerging topic and the scholarship is expected to be in flux, the current version can be viewed as a solid foundation. Congratulations to Metalicat for bring this complex and far-reaching topic to GA family. A.Cython(talk) 20:44, 31 March 2026 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.