| Revision as of 09:49, 20 December 2024 editAlalch E. (talk | contribs)Extended confirmed users, New page reviewers, Rollbackers29,925 edits OneClickArchived "Troy Baker / Voiceverse NFT plagiarism scandal section" to Talk:15.ai/Archive 1← Previous edit |
Revision as of 09:49, 20 December 2024 edit undoAlalch E. (talk | contribs)Extended confirmed users, New page reviewers, Rollbackers29,925 edits OneClickArchived "Disproportionate scrutiny?" to Talk:15.ai/Archive 1Next edit → |
| Line 96: |
Line 96: |
|
::::Oh wait, there already is one. ] ] (]) 04:08, 15 November 2024 (UTC) |
|
::::Oh wait, there already is one. ] ] (]) 04:08, 15 November 2024 (UTC) |
|
{{Archive bottom}} |
|
{{Archive bottom}} |
|
== Disproportionate scrutiny? == |
|
|
|
|
|
Having roamed around other Misplaced Pages articles to edit in my spare time, it feels like the level of scrutiny being applied to this article seems disproportionate compared to many other Misplaced Pages articles. While maintaining high standards is paramount, I notice that articles with far more questionable sources often go unchallenged. The discussion around the Troy Baker/Voiceverse incident involves multiple reliable sources (both primary and secondary) documenting a significant event involving 15.ai, yet faces intense push for removal, while similar incidents documented with comparable sources remain unchallenged in other articles. We should focus on improving content rather than removal, especially when reliable sources exist. Let me know what you think! ] (]) 19:59, 25 November 2024 (UTC) |
|
|
:Its not an {{tq|intense push for removal}}, its just that if you say things that aren't true Wikipedians are possibly the group most likely to correct you. ] (]) 20:59, 25 November 2024 (UTC) |
|
|
::@] I'm sorry, but this isn't a valid justification for undoing the latest revision. You insist that plagiarism never took place ({{tq|also this is probably factually incorrect (the company didnt plagiarize work, someone in the marketing department used an audiofile generated by 15.ai and pretended it was generated via their software}}) despite the reliable sources listed above that literally state that Voiceverse stole work from 15.ai ({{tq|Troy Baker's Partner NFT Company Voiceverse Reportedly '''Steals''' Voice Lines From 15.ai}}, {{tq|Voiceverse NFT admits to '''taking''' voice lines from non-commercial service}}, {{tq|Troy Baker-backed NFT company admits to '''using content without permission'''}}, {{tq|Voiceverse NFT Service Reportedly Uses '''Stolen''' Technology from 15ai}}), and you continue to dismiss the side disagreeing with you. The assertion that Voiceverse plagiarized (and to repeat, Oxford Dictionaries defines "plagiarism" as {{tq|the practice of taking someone else's work or ideas and passing them off as one's own}}) 15.ai is well-documented by multiple sources, and there is no consensus in this talk page whether to keep the section. |
|
|
::I'm going to have to ask you ''why you believe that no plagiarism took place and why it isn't notable when the aforementioned plagiarism is the headline in multiple articles''. In the meantime, I'll revert your edits. As I don't want to start an edit war, I'm hoping that we will come to a consensus in a civil manner. Cheers! ] (]) 21:11, 25 November 2024 (UTC) |
|
|
:::{{ping|GregariousMadness}} See ]. ] (]) 21:12, 25 November 2024 (UTC) |
|
|
:::{{tq|this isn't a valid justification for undoing the latest revision}} if you are referring to that comment then I agree, but I explained my position elsewhere. {{tq|You insist that plagiarism never took place}} The quote you are using shows that I use the word probably, which people who insist on something never do. {{tq|reliable sources listed above that literally state that Voiceverse stole work}} if I create an AI model and you take some of its output while ignoring my ToS, that isn't plagiarism, right? Thats just breaking the ToS. If you steal my actual work, the AI model itself, things might be different. {{tq|there is no consensus in this talk page whether to keep the section}} see ]. {{tq|I'll revert your edits}} you shouldn't because editwarring is uncool and we haven't finished discussing things yet. ] (]) 21:20, 25 November 2024 (UTC) |
|
|
::::Yes, you're right that you used the word "probably", which is why I stopped myself from reverting your edit. I'm hoping that we can come to a consensus about this topic. {{tq|if I create an AI model and you take some of its output while ignoring my ToS, that isn't plagiarism, right? Thats just breaking the ToS. If you steal my actual work, the AI model itself, things might be different.}} No, that's ''both'' stealing/plagiarism ''and'' breaking the TOS. This has happened many times in academia, where students have been expelled for plagiarism simply for misattributing sources, and recent orientation curriculums ''specifically include the misattribution/stealing of generative AI outputs in their research papers as an example of '''plagiarism''''' (they warn all students of this on day 1 of orientation). {{tq|you shouldn't because editwarring is uncool and we haven't finished discussing things yet.}} You're right, and I refrained from reverting your edit. ] (]) 21:27, 25 November 2024 (UTC) |
|
|
::::See, for example: |
|
|
::::{{tq|1. Plagiarism: Copying and pasting text, images, media, etc. generated by AI software into your document without attribution counts as plagiarism as defined by Duke. Repeating or slightly modifying phrases, sentences, or passages generated by AI tools without attribution is also plagiarism. Proper scholarly procedures require that all quoted material be identified by quotation marks or indentation on the page, and the source of information and ideas, if from another, must be identified and be attributed to that source. As described in the Duke Community Standard, plagiarism is not tolerated and may result in disciplinary action. Note that using AI as a grammar correction tool does not count as plagiarism. Duke library has a website with resources on proper AI citation.}} ] (]) 21:29, 25 November 2024 (UTC) |
|
|
:::::I do not believe that the developer of an AI model owns the copyright to all output of that AI model, and certainly not if the AI model was trained on the copyrighted works of others. It would make more sense to give the owners of the copyrights of the works the model is based on the copyright. The law is slow to change. While human authors have copyright, AI models do not at this point. The example of students expelled for plagiarism is, in my view, not comparable because the university cares about whether they've done the work themselves, not if they are infringing on the rights of others. And the university is certainly not trying to protect the hypothetical rights of an AI model. ] (]) 21:31, 25 November 2024 (UTC) |
|
|
::::::While you raise some interesting points about AI copyright law (and I tend to agree with some of the points you make), these are ultimately ''your personal interpretations''. {{tq|The example of students expelled for plagiarism is, in my view, not comparable because the university cares about whether they've done the work themselves, not if they are infringing on the rights of others. And the university is certainly not trying to protect the hypothetical rights of an AI model.}} The comparison to academic plagiarism remains relevant because plagiarism and copyright are distinct concepts - plagiarism is about misrepresenting the source or creator of work, regardless of copyright status. |
|
|
::::::In the Voiceverse case, the key issue isn't about who owns copyright to AI outputs. It's that they explicitly claimed content generated by 15.ai was created by their own different system. That misattribution fits the basic definition of plagiarism as stated in the above PDF: presenting someone else's work as your own. |
|
|
::::::{{tq|I do not believe that the developer of an AI model owns the copyright to all output of that AI model, and certainly not if the AI model was trained on the copyrighted works of others.}} Your view that AI model developers shouldn't own copyright to outputs is a legitimate position in an ongoing debate, ''but it doesn't change the fact that falsely claiming to be the source of content is plagiarism, regardless of copyright status.'' Even if content is in the public domain or has ambiguous copyright, presenting it as your own original work is still plagiarism. ] (]) 21:37, 25 November 2024 (UTC) |
|
|
:::::::We have now reached what I believe to be the crux of the issue, in order to plagiarize the "victim" needs to be the copyright owner of the work. But since the "victim" is an AI model, a glorified Markov chain, it owns nothing and it cannot posses copyright. In the same way you can't steal physical possessions owned by an AI model you also can't steal intellectual possessions owned by an AI model. See ]. I can't plagiarize the sound of the waves, or the wind rustling some leaves. Mother Nature (although personified here) has no legal personhood. Claiming that your AI model generated somehing another AI model generated is possibly not nice, but I don't think it can qualify as plagiarism. As far as I know there is no copyright law against victimless plagiarism. Universities of course do have rules against victimless plagiarism (you are not allowed to pay someone else to do the work for you and then pretend you did it). ] (]) 21:42, 25 November 2024 (UTC) |
|
|
::::::::{{tq|We have now reached what I believe to be the crux of the issue, in order to plagiarize the "victim" needs to be the copyright owner of the work.}} I disagree - the "victim" isn't just the AI model, but the combined set of the AI model and the developer. You wouldn't say you plagiarized Word or Microsoft when you copy someone's text, but you would say you plagiarized the author who wrote using Word. Similarly, 15.ai is both the tool AND the service created by its developer - when Voiceverse used 15.ai's output and claimed it as their own, they were plagiarizing both the system's output and misrepresenting the work of the developers who created that system. |
|
|
::::::::{{tq|But since the "victim" is an AI model, a glorified Markov chain, it owns nothing and it cannot posses copyright. In the same way you can't steal physical possessions owned by an AI model you also can't steal intellectual possessions owned by an AI model. See Personhood.}} The AI model isn't just a "glorified Markov chain" operating in isolation - it's a complete service and system developed by people who specified terms for its use. When Voiceverse claimed output from this system was generated by their own technology, they were misrepresenting both the source of the content (15.ai) and implicitly the work of the developer who created that system. |
|
|
::::::::{{tq|I can't plagiarize the sound of the waves, or the wind rustling some leaves. Claiming that your AI model generated somehing another AI model generated is possibly not nice, but I don't think it can qualify as plagiarism. }} Plagiarism, ultimately, is about honest attribution of sources and work. (And this is what they tell us at orientation, just to be clear.) Just as academic plagiarism can occur with public domain works that have no current copyright holder, plagiarism can occur when misrepresenting the source of AI-generated content, regardless of the complex questions around AI copyright law and who ''technically'' owns what. |
|
|
::::::::The key issue is that Voiceverse claimed someone else's system's output was created by their own different system. That's plagiarism of both the output and misrepresentation of the developer's work, plain and simple. If I were a professor (I'm still a few years away from that, sadly), I would certainly categorize this as a cut-and-dry case of plagiarism - ''who'' was the victim of this plagiarism would be irrelevant. ] (]) 21:52, 25 November 2024 (UTC) |
|
|
:::::::::The law currently does not give an AI model creator the copyright over the output of the AI model. {{tq|the "victim" isn't just the AI model, but the combined set of the AI model and the developer}} not according to the law as I understand it. {{tq|Plagiarism, ultimately, is about honest attribution of sources and work.}} that is not how the law works. For example, you can freely copy text from works in the public domain whose copyrights have expired without attribution. {{tq|If I were a professor}} thank god you are not studying law! ] (]) 21:57, 25 November 2024 (UTC) |
|
|
::::::::::{{tq|The law currently does not give an AI model creator the copyright over the output of the AI model.}} Plagiarism and copyright infringement are two distinct concepts that you're conflating. You're correct about copyright law, but incorrect in using that to argue there was no plagiarism. You're right that current law doesn't give AI model creators copyright over AI output; this is a correct statement about copyright law. '''However, plagiarism is NOT defined by copyright law.''' Plagiarism is about presenting someone else's work as your own, ''regardless of copyright status.'' To reference your example, '''Students must cite public domain works even though they're not copyrighted.''' Academics must acknowledge their sources even when quoting copyright-expired materials. And finally, ''presenting someone else's public domain work as your own original creation is still plagiarism''. |
|
|
::::::::::{{tq|thank god you are not studying law!}} Seriously? That was uncalled for. ] (]) 22:03, 25 November 2024 (UTC) |
|
|
:::::::::::{{tq|Plagiarism is not illegal in the United States in most situations. Instead it is considered a violation of honor or ethics codes and can result in disciplinary action from a person’s school or workplace. However, plagiarism can warrant legal action if it infringes upon the original author’s copyright...}} If there is no "original author", because the AI model has whitewashed copyrighted stuff from other sources and can't itself be an author, how can it be plagiarism? Not sure how you interpreted that, I was just joking about the fact that the law is confusing and weird and less fun than AI. ] (]) 22:06, 25 November 2024 (UTC) |
|
|
::::::::::::{{tq|Plagiarism is not illegal in the United States in most situations. Instead it is considered a violation of honor or ethics codes and can result in disciplinary action from a person’s school or workplace. However, plagiarism can warrant legal action if it infringes upon the original author’s copyright... }} This only strengthens the argument that Voiceverse committed plagiarism, even while we debate the copyright implications. I'm glad we agree that plagiarism is an ''ethical violation'', not a legal one. Copyright infringement is a separate legal issue that may or may not accompany plagiarism; therefore, ''whether 15.ai has legal copyright over its outputs is irrelevant to whether Voiceverse committed plagiarism''. They could still have committed plagiarism (an ethical violation) even if no copyright was infringed. |
|
|
::::::::::::{{tq|Not sure how you interpreted that, I was just joking about the fact that the law is confusing and weird and less fun than AI.}} Not a very funny joke. Students like me spend over a decade working to become a professor in their area of expertise, and I don't like it when someone I don't know on the Internet jokes about my credentials. ] (]) 22:13, 25 November 2024 (UTC) |
|
|
:::::::::::::But I didn't joke about your or anyone else's credentials? ] (]) 22:14, 25 November 2024 (UTC) |
|
|
::::::::::::::You wrote {{tq|thank god you are not studying law!}} on my mention that I am working to become a professor. You don't see how that sounds insulting? ] (]) 22:22, 25 November 2024 (UTC) |
|
|
:::::::::::::::No, that does not sound insulting at all to me. Can you explain why you react this way? ] (]) 22:32, 25 November 2024 (UTC) |
|
|
::::::::::::::::If you truly didn't mean any offense, then I'll consider it water under the bridge. For most of this talk page discussion, it felt like people weren't being friendly towards me (made worse because of a prior interaction on this article), so I took that statement to imply that I was so stupid in my arguments that studying the law would have been a disservice to the law. ] (]) 22:37, 25 November 2024 (UTC) |
|
|
:::::::::::::::::Oh man I apologize if I come across ''that'' dickish, but that was not what I meant at all! I just think studying law is like bashing your head against a wall. If people are unfriendly what usually helps is ] and ]. And sometimes an admin dropping by to remind everyone how to behave helps. ] (]) 22:39, 25 November 2024 (UTC) |
|
|
:::::::::::::::Those do not correlate. There are professors for everything under the sun, no one here had any way to know you were studying to be a law professor until you just now brought it up. — <b>]:<sup>]</sup></b> 22:24, 25 November 2024 (UTC) |
|
|
::::::::::::::::I don't know what you're talking about. I'm not studying to be a law professor... I'm a machine learning grad student working to be a professor, and professors and colleges are the ones who decide whether a student has been plagiarizing or not, not the law. May I ask, why are you confusing the conversation all of a sudden? ] (]) 22:32, 25 November 2024 (UTC) |
|
|
:::::::::::::::::I'm confused because you're not being clear. You're really pissed off over a law joke, when it's got absolutely nothing to do with your studies? |
|
|
:::::::::::::::::Also, I'm ''really'' getting sick of your failure to ] and making accusatory statements towards others. — <b>]:<sup>]</sup></b> 23:13, 25 November 2024 (UTC) |
|
|
::::::::::::::::::<s>Both parties were clear enough, @] and @] had already cleared up the confusion and made up before your unnecessarily confusing and aggressive statements. "You're really pissed off over a law joke", and "Also, I'm really getting sick of your failure to WP:AGF and making accusatory statements towards others." is not civil behavior. ] (]) 01:03, 26 November 2024 (UTC)</s> {{sockvote|HackerKnownAs}} |
|
|
:::::::::::::::::::Says the ]. Regardless, I'm done with this discussion. There's no consensus for adding this to the article, and you can attempt an ] if you really want to add it. — <b>]:<sup>]</sup></b> 17:19, 26 November 2024 (UTC) |
|
|
::::::::::::::::::Please calm down and remember to be civil. We (Polygnotus and I) already cleared up the confusion and are now attempting to reach a compromise before you commented, so this abrasiveness is unnecessary. ] (]) 19:15, 26 November 2024 (UTC) |
|
|
::::::::::::::::I had read their userpage so I knew they were ''not'' studying to become a law professor. ] (]) 22:32, 25 November 2024 (UTC) |
|
|
::::::::::::::::So if we agree that there is legally nothing wrong with what they did, and it was only an "ethical violation" by someone working in marketing, then its a non-story right? Storm in a teacup. Ethical violations and marketing go hand in hand, right? ] (]) 22:36, 25 November 2024 (UTC) |
|
|
:::::::::::::::::{{tq|So if we agree that there is legally nothing wrong with what they did, and it was only an "ethical violation" by someone working in marketing, then its a non-story right?}} No, absolutely not. As a personal aside, ethical violations in marketing shouldn't just be dismissed as normal or acceptable just because they're common. But on top of that, this wasn't just a minor marketing mishap/strategy. Voiceverse misrepresented their technological capabilities to potential investors/customers, used another company's technology while claiming it was their own, did this during a major partnership announcement with Troy Baker, and violated terms of service in the process. The fact that something might not be illegal doesn't make it a "non-story." Again, the notability of the story is supported by multiple sources directly implicating Voiceverse in the plagiarism of 15.ai's work, which should be the only thing that matters. ] (]) 22:42, 25 November 2024 (UTC) |
|
|
::::::::::::::::::In my view those are a lot of reasons why this is important to mention on an article about Voiceverse, but not 15.ai. What impact did it have on 15.ai? Perhaps we can compromise on something much shorter and less prominent that mentions the impact of the plagiarism? It may be possible to get consensus to include something like that. ] (]) 22:46, 25 November 2024 (UTC) |
|
|
:::::::::::::::::::I don't mean to be dismissive, but I have to go soon. But being the victim of plagiarism is absolutely notable for 15.ai, as the incident gained significant media attention specifically because it involved 15.ai being exploited by a larger commercial entity, making it a defining moment in its history. |
|
|
:::::::::::::::::::I'll give a personal analogy: if one of my academic articles was plagiarized by a more important researcher or publisher, it would have a ''significant'' impact on myself and my "history". It would definitely be something that, if I had my own Misplaced Pages article in the future, would be mentioned as an aside. ] (]) 22:50, 25 November 2024 (UTC) |
|
|
::::::::::::::::::::It is always a bit sad when David's Misplaced Pages article is defined by Goliath's actions imo. I think there was previously an undue amount of focus on something that, in my view, was not very important, but if you disagree you may have some suggestions how to incorporate it into the article in a less prominent way. ] (]) 22:53, 25 November 2024 (UTC) |
|
|
:::::::::::::::::::::Thank you for the engaging conversation yesterday. Apologies for jumping the gun on your joke - I was feeling antsy because this is my first time engaging in a serious debate like this on Misplaced Pages, and I'm not very good at serious debates. I took some time to think about it myself, and as you suggested, I'll be working on an edit that addresses your concerns while not making the event a centerpiece of the article. I'll let you know once I finish my edits! ] (]) 19:14, 26 November 2024 (UTC) |
|
|
:::::::::::::::::::::I've significantly trimmed down the section to only include a paragraph and a half of text. I'm thinking of trimming it down even further, but I believe it no longer puts an undue amount of focus on the event. Let me know if I should change anything further! Cheers. ] (]) 19:30, 26 November 2024 (UTC) |
|
|
:::::::::::::::::::<s>I agree with a compromise similar to this - it was what I suggested earlier, so I'm happy we're coming to a consensus! We can still include the citations but condense the section a little bit. I'm sure we can find a good middle ground. ] (]) 01:14, 26 November 2024 (UTC)</s> {{sockvote|HackerKnownAs}} |
|
|
:There's nothing unusual here, especially for an article that recently had multiple ] issues. As to {{tq| I notice that articles with far more questionable sources often go unchallenged}}, that's an ] argument. Feel free to bring up the problems those articles have so they get some attention. It's a volunteer project, so people pick and choose what they get involved with. Until they know there's a problem, they can't to do anything about it. — <b>]:<sup>]</sup></b> 21:13, 25 November 2024 (UTC) |
|
|
::I see. I was beginning to think there was a coordinated effort to remove content from an article that I drafted up as my first Wikipedian project, but that makes more sense. I still think that the contentious section has more than enough references and relevant to keep, but I understand the scrutiny now. ] (]) 21:19, 25 November 2024 (UTC) |
|
|
:::Please list or tag problems in other articles, and if you ping me I'd be happy to take a look. ] (]) 21:21, 25 November 2024 (UTC) |
|
|
::::Sure. One article that I recently looked at that has a ''lot'' of unsourced text is ], but having been a participant in Olympiad mathematics, I can safely say that it is notable and much of the information there is correct. If we followed the standards exhibited in this talk page, however, I would feel like over 70% of the article would have to be removed. The 15.ai article is extremely well-sourced in comparison. ] (]) 21:54, 25 November 2024 (UTC) |
|
|
:::::Thank you. I have added the {{tl|Sources}} template which adds it to ]. ] (]) 22:01, 25 November 2024 (UTC) |
|
|
:{{tq|We should focus on improving content rather than removal}} |
|
|
:Removing content '''''is''''' sometimes the same as improving the article. Mentions of a scandal, complete with tweets blasted across the article, does not improve the encyclopedic content of the article. It isn't a major event for 15.ai, the topic of the article, and it wasn't a prolonged controversy with sustained converage that warrants anything more than a brief mention. An entire section of the article about it is ]. Moreover, ] states {{tq|Information should not be included solely because it is true or useful. An article should not be a complete presentation of all possible details, but a summary of accepted knowledge regarding its subject}}. A brief scandal of another company using a voiceline that was generated by 15.ai has functionally very little to do with 15.ai itself. It is, essentially, a conflict between 15(the creator of 15.ai), who for the purposes of Misplaced Pages is considered unnotable, , and Voiceverse, who for the purposes of Misplaced Pages is also considered unnotable. <b>]</b> 08:25, 26 November 2024 (UTC) |
|
|
|
|
|
== Regarding The Batch == |
|
== Regarding The Batch == |
|
|
|
|
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.
I wonder if this RfC is still needed. I haven't followed the dispute but it looks to me like it might only exist because a sock-farm was trying to keep option A and with this sock farm now hopefully gone it might be unnecessary. I appreciate one editor is supporting B above while the rest are supporting C and maybe the RfC having started it's too late but it just looks to me like the sort of thing which could have been resolved via normal discussion were it not for the sock farm. Nil Einne (talk) 17:40, 17 November 2024 (UTC)
I have been removing all instances of this source from the article. It was deemed unreliable when the article was a draft. They were re-added after the article was released from draftspace by a now vanished user who flagged them as a minor edit. Brocade River Poems (She/They) 00:02, 28 November 2024 (UTC)
I'm posting this as a reference. According to archived discussions on the Teahouse, a good rule of thumb is three independent and reliable sources that demonstrate significant coverage to establish WP:GNG. Here are the major reliable sources that provide significant coverage for the subject:
Yes, I'm aware that this is a contentious article that was submitted to AfD. But the AfD was closed largely due to the misbehaving of new Misplaced Pages editors, who are likely to be children, which is not surprising given the popularity of the application among younger people. I'm committed to doing this subject justice, and I argue that this subject not only meets the bare minimum of notability, but meets it well-within question. GregariousMadness (talk to me!) 17:11, 19 December 2024 (UTC)
