Wikipedia:Good article reassessment/15.ai/1

15.ai

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 • Watch article reassessment page • Most recent review

Result: No consensus. ~~ AirshipJungleman29 (talk) 17:23, 5 November 2024 (UTC)

Detailed concerns about large-scale COI editing and neutrality issues in this article can be found at Talk:15.ai#Concerns about this article. GA criterion 4 is thus under serious question. ~~ AirshipJungleman29 (talk) 09:57, 23 October 2024 (UTC)

The vast majority of the current iteration of the article was as it was when the article was first accepted as a good article years ago. Much of the COI edits have been from vandals, and it's evident that the article has a major vandalism problem, the subject being a rather popular topic of discussion. I disagree with removing the GA label due to the edits of some bad-faith actors. HackerKnownAs (talk) 23:42, 23 October 2024 (UTC)

While is this a good point, I would also like to express the concern over the fact that many of the citations make heavy references to one Kotaku article. Although I am repeating an concern already expressed in the above talk topic, I feel that it should be mentioned in this talk topic as well. Thought 1915 (talk) 00:54, 24 October 2024 (UTC)

I will note, as I did on the talk-page discussion linked, that the GA approval very probably should have never happened. The original COI discussion topic was never rectified and was randomly expunged from the talkpage before the article was assessed.[1] The individual who assessed the article as good only assessed one other article as good, and that article was deleted for large amounts of copyvio. It was an assessment provided by a new reviewer, after an IP editor randomly elevated the article to B Status[2] and expunged an entire talkpage discussion about COI editing. The one who reviewed the article has done little else Special:Contributions/SirGallantThe4th after the review, departed, and returned only to defend 15.ai from deletion[3] and then promptly returned to the ether. Again, there was considerable activity by individuals involved in the "Pony Preservation Project" to edit this article [4] including bragging about having their artwork featured on a wikipedia article [5] the same artwork that is presumably the logo that was copyvio'd off the wikimedia commons and which was improperly re-added to the article. There are blatant references on the PPP thread on the archive _which was included as a source_ on the article that show users suggesting fabricating sources and showing a coordinated effort to drive the direction of the article. Given the involvement of the PPP with 15.ai and the extensive editing done by members of the PPP, it is clear WP:COI. Given the circular nature of the sources (which were deemed reliable per Wikipedia:WikiProject_Video_games/Sources#Reliable_sources, which is dubiously applicable to this article, as the article is tangentially related to video games) and the fact that Wikipedia:WikiProject_Video_games/Sources#Reliable_sources says that editors should be cautious about blog/geeky posts that have little news or reporting significance which the 15.ai article clearly is. If the Kotaku isn't reliable as "blog/geeky posts", that means every source that references the Kotaku article is likewise unreliable. Given that, the WP:NOTABILITY of the article itself is dubious. There are sources from Jan 2021 and then it only resurfaces in media covering a controversy of Voicesense, not articles about 15.ai itself. The problem with this article goes much deeper than some bad-faith actors who made random edits to the article over the years. Brocade River Poems (She/They) 05:43, 24 October 2024 (UTC)

Per WP:RGA, there's no minimum experience requirement for GA reviewers – questioning an editor's review history isn't based on Wikipedia policy.

The COI claims don't meet WP:COI burden of evidence – community discussion on PPP forums doesn't constitute "direct financial or close personal relationships."

Per WP:VG/RS, gaming outlets like Kotaku are considered reliable for tech coverage. The article meets WP:GNG through significant coverage from multiple reliable sources. Gaps in coverage don't invalidate WP:SUSTAINED once notability is established.

Removing talk page content followed WP:TPO guidelines for outdated discussions. If specific violations exist, they should be raised at WP:AN/I rather than used to challenge GA status.

Furthermore, the last time WP:NOTABILITY of the subject was questioned, it was unanimously agreed that it met notability. Per WP:NTEMP, the relevant quote is "Notability is not temporary; once a topic has been the subject of "significant coverage" in accordance with the general notability guideline, it does not need to have ongoing coverage." HackerKnownAs (talk) 05:08, 25 October 2024 (UTC)

Per WP:RGA, there's no minimum experience requirement for GA reviewers – questioning an editor's review history isn't based on Wikipedia policy.

It is entirely relevant when the editor in question had little activity on Wikipedia, reviewed two articles and two articles alone, and flagged them both as good when both articles had issues that should have failed them. Very specifically, this article failed GA Criteria for reliable sourcing when it was reviewed because it used WeGotThisCovered as a source [6] which has been an unreliable source on Wikipedia since 2020 WP:WEGOTTHISCOVERED as well as using The Batch, which during the draft process was said to be an unreliable source.

The COI claims don't meet WP:COI burden of evidence – community discussion on PPP forums doesn't constitute "direct financial or close personal relationships."

The COI claims that were first brought against yourself, PortalFan22, and GregariousMadness Talk:15.ai#COI were never adequately addressed and were perfectly relevant. Moreover, the discussion was wrongfully expunged by IP Editor vandalismSpecial:Diff/1090463388. When the project you are writing about contains the notation Special shoutouts go to 4chan's /mlp/ and its anons who have spent hundreds of hours collecting, cleaning, and organizing clips of dialogue taken from the show My Little Pony: Friendship Is Magic. Their collective efforts as well as their constructive criticism via thorough tests of experimental model versions have proven to be extremely helpful and ultimately indispensable to the development of my work and when the PPP cites 15.ai's code of conduct as their own code of conduct, yes, participating in off-wiki discussions about editing an article about something they are heavily involved in the development of constitutes a WP:COI. When material goods, such as artwork and logos, are being provided to 15.ai by PPP, that is a clear relationship. When PPP is directing individuals to edit the article to include information about PPP that is a clear and blatant conflict of interest.

Per WP:VG/RS, gaming outlets like Kotaku are considered reliable for tech coverage.

Per WP:VG/RS, News posts from Kotaku between 2010 and 2022 are considered reliable, although editors are cautioned of blog/geeky posts that have little news or reporting significance. The Kotaku Article is filed under Odds and Ends [7], not News. The article is not tagged as News. It is not a news post.

The article meets WP:GNG through significant coverage from multiple reliable sources. Gaps in coverage don't invalidate WP:SUSTAINED once notability is established.

Furthermore, the last time WP:NOTABILITY of the subject was questioned, it was unanimously agreed that it met notability.

Notability isn't established. The coverage is trivial, and the previous AfD on this article was interfered with by blatant WP:SPA accounts that accounted for 4 Keep votes[8][9][10][11], and the AfD was specifically closed with the message Although not unanimous. The article plainly fails WP:SIGCOV and WP:NSUSTAINED which says Brief bursts of news coverage may not sufficiently demonstrate notability. The Eurogamer coverage isn't about 15.ai, the 2022 converage is solely about the Voiceverse controversy and stolen content from 15.ai, it is not sufficient to establish notability for 15.ai.

I find it odd that you consider an IP Editor who did nothing but vandalize Wikipedia pages aside from increasing this article to B Status and expunging a thread about WP:COI editing to have followed WP:TPO guidelines. In fact, WP:TPO explicitly states: The basic rule, with exceptions outlined below, is to not edit or delete others' posts without their permission. The only exception given for deleting talk page content is Delete. It is common to simply delete gibberish, test edits, harmful or prohibited material (as described above), and comments or discussion clearly about the article's subject itself (as opposed to comments and discussion about the treatment of the subject in the article).

Nothing in WP:TPO allows for the wholesale deletion of a valid concern of WP:COI just because 3 months had transpired since the issue was brought up. Archival exists for a reason. Closing discussions exists for a reason. It is wholly inappropriate to delete and expunge the COI topic. Brocade River Poems (She/They) 12:44, 25 October 2024 (UTC)

With all due respect, the suggestion that 15.ai does not meet notability is dumbfounding. It pioneered accessible neural voice synthesis, was widely covered in tech media, and influenced numerous subsequent AI voice projects. I would not be exaggerating when I say its advent was one of the biggest news in the AI space in 2020 and 2021. Whether or not you agree with how the GA review was conducted, the project’s significance in AI development is well-documented and indisputable. The coverage is certainly not trivial, and the attempt to downplay it by cherry-picking coverage gaps or questioning the reliability of established tech journalism (and patently ignoring the sources that *are* reliable) does not seem motivated by Wikipedia policy.

~~ SirGallantThe4th (talk) 17:28, 25 October 2024 (UTC)

With all due respect, your continued penchant of vanishing from Wikipedia and returning only for championing the existence of this article is highly unusual.

I would not be exaggerating when I say its advent was one of the biggest news in the AI space in 2020 and 2021 the project’s significance in AI development is well-documented and indisputable

These are not only wholly uncited and unsubstantiated claims, but claims which would indicate you have some sort of interest in this project and, again, should have never been involved in assessing the article in the first place. I am not going to repeat my points about why much of the sourcing is unreliable. If there is such wide, well-documented, and indisputable source coverage, where is it? Why is none of it represented in the article? Why is it that the only sources used were WP:SPS and unreliable sources?

The coverage is certainly not trivial, and the attempt to downplay it by cherry-picking coverage gaps or questioning the reliability of established tech journalism (and patently ignoring the sources that *are* reliable) does not seem motivated by Wikipedia policy.

It's not "cherry-picking" to state the plain fact that the sources used are not reliable for the purposes they are being used for. Kotaku's Odds and Ends culture section is not News. Kotaku is deemed reliable for News Posts during this time period. Articles which circularly refer to Kotaku's Odds and Ends piece within the same week as the Kotaku article does not represent WP:SIGCOV. Brocade River Poems (She/They) 23:13, 25 October 2024 (UTC)

What a concerning post. It’s extremely strange that my editing patterns are being used to cast suspicion on the subject’s notability and its GA status. As someone who works in AI, I naturally took interest in reviewing the 15.ai article given its significance in voice synthesis - in its heyday, it was literally the biggest thing in the voice AI space. The implication that my “return” to defend it from deletion was suspicious ignores the simple fact that many editors follow topics they find an interest in, and the fact that one can stay logged into my account without wanting to contribute to Wikipedia.

I suggest stepping back and seeing how presumptuous (and frankly alienating) your comments are. You’ve crafted an elaborate theory about coordinated editing and suspicious motives based solely on contribution patterns. Not every editor needs to be constantly active to make valid contributions, and returning to defend an article I reviewed from deletion is perfectly natural. Occam’s Razor applies here, and I hope anyone else who reads this can see it for themselves as well.

~~ SirGallantThe4th (talk) 17:18, 25 October 2024 (UTC)

There appears to be coordinated disruptive editing going on. While the 15.ai article has experienced vandalism before, the current situation is unprecedented. The above editor's removal of well-sourced content followed by claims of insufficient citations is particularly concerning. That a non-neutral paragraph that blatantly violated WP:YESPOV (which I reverted recently) was "approved at DRN" is also strongly questionable, especially given the unusual spike in DRN activity as well.

I plan to restore the article next week in accordance with WP:BRD, and I hope other long-term editors familiar with the article's development can assist (including yourself). Per WP:ATD, I don't believe GA status should be removed until we've made a good-faith effort to address any legitimate concerns through collaborative editing. The current issues, while worth discussing, can be resolved through normal Wikipedia processes rather than immediately reverting to non-GA status. HackerKnownAs (talk) 18:15, 25 October 2024 (UTC)

I suggest stepping back and seeing how presumptuous (and frankly alienating) your comments are.

There is broad evidence of off-Wiki coordination to edit the article, coupled with the emergence of clear WP:SPA activity and manipulation in both the AfD and the editing of the article. It is not, frankly, presumptuous or absurd to suspect something is suspicious about an editor who erroneously assesses 2 articles as good, one of which is full of copyvio, and then disappears for an extended amount of time and returns only to defend this article. Brocade River Poems (She/They) 23:19, 25 October 2024 (UTC)

You’ve said multiple times that you have evidence of off-wiki coordination to edit the article, and yet you haven’t posted solid evidence at all. I wasn’t aware of any forum before I came across this article while surfing through random AI related articles two years ago, so how does that make my involvement with the article a coordinated affair? The burden of proof lies with the editor making serious accusations, and forum discussions and editing patterns alone don’t constitute evidence whatsoever. If you have actual evidence of coordination, please post them instead of throwing unfounded accusations at editors who just want to help.

~~ SirGallantThe4th (talk) 04:06, 26 October 2024 (UTC)

It would be bludgeoning and belaboring the point to reiterate, again, the issues I already pointed to in the talk-page discussion and above, if you are not inclined to read what is already provided, I cannot help you. Whether you yourself were involved in the coordination is immaterial, my point is that because there was demonstrable coordination it is not unreasonable to view your assesment, disappearance, and return solely to defend the article, subsequent re-disappearance, and subsequent re-return to defend the article, with suspicion given the fact that the AfD was manipulated. I have levied no specific accusation against you beyond the fact that your assesment was improper, that your only other assesment was deleted for copyvio, that you were an inexperienced editor, and that you did little else after the review was done. All of these statements are easily verifiable. After your approval of 15.ai you made 9 Talk Page Edits, 7 Edits flagged as minor, and 8 Mainspace edits and then you disappeared. If you feel particularly aggreived at my characterization of your activity as "little" prior to reviewing the articles, I quite specifically mean that within your first 25 edits on Wikipedia you assessed an article now deleted for copyvio as good and then assessed an article that utilized a source deemed unreliable since 2020 and sources deemed unreliable when the article was a draft as good. That isn't a whole lot of activity and represents a lack of experience. Your improper assesment of the article as Good was also used at the AfD with the discussion being relisted by an admin who commented I'm very reluctant to delete an article that is a current GA. Brocade River Poems (She/They) 12:49, 26 October 2024 (UTC)

I would also like to point out that I haven’t done “little else” on Wikipedia, I was responsible for creating the Berlin draw section in the Berlin Defence article, which has since been heavily cited numerous times in Chess.com articles, Youtube videos, and other places on the Internet. It’s not fair to pass judgment for not being active on Wikipedia when I prefer to edit articles where I’m familiar with the subject rather than editing as many articles as possible to pump my contributions number up.

If you want more volunteers to help improve Wikipedia, I suggest not scrutinizing casual editors. I’m honestly a bit offended that I somehow have less credibility just because I don’t edit articles frequently.

~~ SirGallantThe4th (talk) 18:02, 25 October 2024 (UTC)

I was responsible for creating the Berlin draw section in the Berlin Defence article

Which you did before the assesment. My statement, which you are apparently misunderstanding, was The one who reviewed the article has done little else...after the review. Which is to say, you made few edits after you assesed the article and then you left for 6 months and returned only for the AfD and then departed again. Brocade River Poems (She/They) 12:58, 26 October 2024 (UTC)

I wasn’t aware editors were required to edit Wikipedia 24/7 to have their past contributions count. It doesn’t look like it from reading WP:VOLUNTEER. Apologies for having a life outside Wikipedia. I’ll try to do better.

~~ SirGallantThe4th (talk) 15:44, 26 October 2024 (UTC)

Ignoring the maelstrom above, much of which seems to have little to do with the GA criteria (notability concerns should be taken to WP:AFD, conduct concerns to WP:ANI, COI concerns to WP:COIN), the point of relevant contention seems to be whether the article is overdependent on a Kotaku source? ~~ AirshipJungleman29 (talk) 14:40, 28 October 2024 (UTC)

The main point of contention for the Good Article Criteria is reliability of sources and breadth of coverage. The Kotaku used isn't from the part of Kotaku that Wikipedia:WikiProject_Video_games/Sources#Reliable_sources clarifies is generally reliable. The Kotaku used comes from Kotaku's "Odds and Ends" section, with VGRSN noting that caution should be applied to Kotaku's geeky/blog content, and all of the coverage of 15.ai comes either the same way of the Kotaku article, almost always referncing the Kotaku Article, or comes a year later with trivial mentions of 15.ai while covering the Voiceuniverse plagarism controversy. Prior to my excising them, the article relied heavily on WP:SPS and stuff posted on Gwern and other personal blogs. Gwern, notably, describes itself as The goal of these pages is not to be a model of concision, maximizing entertainment value per word, or to preach to a choir by elegantly repeating a conclusion. Rather, I am attempting to explain things to my future self, who is intelligent and interested, but has forgotten as well as fansites such as "Equestria Daily", which have been previously noted at WP:RSN to generally be reliable when dealing with interviews of the cast/crew of My Little Pony or official coverage from Hasbro only Wikipedia:Reliable_sources/Noticeboard/Archive_100#Two_My_Little_Pony:_Frienship_is_Magic_sources.

In terms of "Broad Coverage", it stays focused on the topic without going into unnecessary detail, the article does not seem to stay focused on 15.ai and a number of the sources used and a great deal of the content is simply about the underlying technology and a scandal.

The neutrality of the article is also up for debate as the "Reception" section includes only positive mentions and sources misrepresented to praise 15.ai specifically. For instance, Rock, Paper, Shotgun that is cited says Machine learning is absolutely fascinating and yet I mostly just enjoy when people use impressive tech to create weird skits and memes, the article currently represents this statement in the reception section as Lauren Morton of Rock, Paper, Shotgun and Natalie Clayton of PCGamer called it "fascinating,". Similarly, the PCGamer piece actually reads Spotted by Kotaku over the weekend, 15.ai is a deep-learning text-to-speech tool trained on a library of audio clips for dozens of characters. It's all very fascinating to read about These quotes have been misconstrued into being glowing reviews of 15.ai itself when the articles are simply saying that the underlying technology is fascinating or fascinating to read about. Brocade River Poems (She/They) 01:27, 30 October 2024 (UTC)

Are you able to fix these issues BrocadeRiverPoems? ~~ AirshipJungleman29 (talk) 15:23, 1 November 2024 (UTC)

this entire discussion sounds more like one person with a personnal vendetta against it

my opinion is there are alot of articles that have way worse sources that people should focus on deleting, not this one 108.191.41.11 (talk) 00:45, 31 October 2024 (UTC)

Deletion of this article is not being proposed here. Trainsandotherthings (talk) 22:32, 1 November 2024 (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.