Talk:Artificial intelligence art: Difference between revisions
Camdoodlebop (talk | contribs) →Adding a Stable Diffusion example: new section |
|||
Line 76: | Line 76: | ||
There has been a recent introduction of an unsourced, false sentence saying "There are many other AI art generation programs, '''''ranging''''' in complexity from simple consumer-facing mobile apps to Jupyter notebooks ..." This is an unsourced and false description that AI programs only exist in this very narrow range, since there is of course a long history of AI artists and researchers who are always creating and using new programs that are not available to the general public. There are many types of AI that are simpler than a mobile phone app, and many types of AI that are more complex than a Jupyter notebook. I previously edited this sentence to instead say "There are many other AI art generation programs, '''''including''''' ..." so as not to give the false impression that there are no programs outside of this very limited range. This change was undone by editor Camdoodlebop. The edit summary they used was "reverted vandalism." I am about to remove this false, unsourced "ranging" description again, and I am leaving this talk page message in case anyone wants to discuss whether they have a source saying AI only exists in this very narrow range or whether my correcting this false statement is "vandalism." [[User:Elspea756|Elspea756]] ([[User talk:Elspea756|talk]]) 13:15, 1 September 2022 (UTC) |
There has been a recent introduction of an unsourced, false sentence saying "There are many other AI art generation programs, '''''ranging''''' in complexity from simple consumer-facing mobile apps to Jupyter notebooks ..." This is an unsourced and false description that AI programs only exist in this very narrow range, since there is of course a long history of AI artists and researchers who are always creating and using new programs that are not available to the general public. There are many types of AI that are simpler than a mobile phone app, and many types of AI that are more complex than a Jupyter notebook. I previously edited this sentence to instead say "There are many other AI art generation programs, '''''including''''' ..." so as not to give the false impression that there are no programs outside of this very limited range. This change was undone by editor Camdoodlebop. The edit summary they used was "reverted vandalism." I am about to remove this false, unsourced "ranging" description again, and I am leaving this talk page message in case anyone wants to discuss whether they have a source saying AI only exists in this very narrow range or whether my correcting this false statement is "vandalism." [[User:Elspea756|Elspea756]] ([[User talk:Elspea756|talk]]) 13:15, 1 September 2022 (UTC) |
||
== Adding a Stable Diffusion example == |
|||
Would it be acceptable to add an example from Stable Diffusion that a user created themselves, or would they need to source the example from a third party to avoid self-promotion? I have some examples I've uploaded to Wikimedia I could add. [[User:Camdoodlebop|Camdoodlebop]] ([[User talk:Camdoodlebop|talk]]) 00:41, 10 September 2022 (UTC) |
Revision as of 00:41, 10 September 2022
Self promotions!
People seriously wikipedia is not a place to self promote or promote your website! Please stick to the facts! — Preceding unsigned comment added by Norttis (talk • contribs) 19:56, 22 February 2021 (UTC)
Edit Warring/Reliable Source Discussion
Please be careful about running into edit warring. An edit has been made, and then reverted. The next stage is discussion - the burden of evidence is on DaveOomah, as they are the one trying to add new content, and need to convince that it is a viable edit and not, in fact, promotional. While not enough on its own, a critical part of this is going to be demonstrating that Coinpost is a sufficiently reliable and independent publication. Comments like "Even if you apply from a press release, if the CoinPost editorial department determines that it is important information, it will be adopted as news." (on their press release page) is somewhat concerning, as it suggests that insufficient editorial control and fact-checking is being exerted, as well as independence issues. However, perhaps I can be proved wrong. If you guys can't agree, then you can raise the question on reliable sources noticeboard. Nosebagbear (talk) 14:44, 9 April 2021 (UTC)
Repeated additions of false self-promotional claim of "world's first published AI Art book"
This article contains the obviously false claim that in 2019, the "world's first published AI Art book" was published, and includes a promotional image. I have tried to remove this, however "Norttis" has repeatedly added this obviously false information and spam image. "Norttis" has also removed references to AI art projects from the 1960s through 1990s, in an apparent attempt to bolster their false claim the first AI book was published in 2019, over 50 years after several prominent AI art projects. "Norttis", do you have any defense for including this obviously false information? And do any other editors have an opinion on this? Elspea756 (talk) 17:44, 30 July 2021 (UTC)
- You seem to have no clear understanding of AI art is and what computer generated art is. A generative adversarial network (GAN) is a class of machine learning frameworks designed by Ian Goodfellow and his colleagues in 2014, so what ever happened in the 1960's is not relevant on this page.
- I created this page over 2 years and have kept it as relevant and up to date as I can, and many others have added to it. In a single day you take it upon yourself to rewrite almost the entire page, not even bothering to ask anything in the talk section before doing so. You removed not only my work but the work may other wikipedians, and tried to turn the page into what you wanted, not what it should be. — Preceding unsigned comment added by Norttis (talk • contribs)
- The topic of this article is not "generative adversarial network art". Any art produced by AI, even in the 1960s, would be on topic. - MrOllie (talk) 18:27, 30 July 2021 (UTC)
- There have been efforts to apply machine learning techniques such as genetic computation to art since well before anyone thought of GANs. Also, please remember to sign your posts. - MrOllie (talk) 18:33, 30 July 2021 (UTC)
AI has existed as an academic field since 1956. Yes, in the 60s, they used very different techniques than AI does today. Not because those techniques were not "artificial intelligence", but because those techniques had limits. From a historical perspective, it would be a smart bet to guess that current techniques will also have limits, and that new techniques will be found. Thus, we, as Wikipedians, can't just redefine the term AI to mean "AI in the 2020s". AI is whatever people called AI. It's just one field of study, with a history.
If someone used a technique to create art that came from AI's huge arsenal, in any decade, then we would have to label it as "AI art". Even if they used AI techniques invented in the 50s (such as adversarial search, or generative grammar, or neural networks). Or a combination of these (which is what "generative adversarial network" sounds like).
Computer art, by the way, is a term used by the world to describe any art that involves a computer at some stage -- even photoshop collages. ---- CharlesGillingham (talk) 06:05, 19 August 2021 (UTC)
Attempted removal of Harold Cohen, etc. yet again
Editor "Nortis" has been repeatedly warned against removing early examples of AI art by multiple editors including myself, Mr. Ollie, and Charles Gillingham above on this talk page. "Nortis" has returned to delete every example of AI artists because those examples apparently disprove Nortis's theory that AI art began just a couple of years ago. For example, Nortis has particularly targeted artists like Harold Cohen, whose work is described as "AARON was an artificial intelligence program created by Harold Cohen (1928–2016), artist and pioneer in computer-generated art" by https://direct.mit.edu/leon/article/54/4/412/97258/Harold-Cohen-and-AARON-Collaborations-in-the-Last Nortis has removed this and other examples with the summary "Listing random people in the field is just nonsense" and "A computer programme is not AI art." And we also now have "Drmies" also removing this information, with their justification being that mentioning Harold Cohen and other AI artists in this article is "namedropping, promotional sourcing." So, let's reopen this discussion on whether AI artists like Harold Cohen should be discussed in this article on AI art. Thank you. -- Elspea756 (talk) 02:41, 14 January 2022 (UTC)
- I don't see any consensus on this talk page or in the article history to bloat this article with an arbitrary list of names of people loosely associated with this topic; that's what categories are for. OhNoitsJamie Talk 03:54, 14 January 2022 (UTC)
- Thank you for joining this conversation, OhNoitsJamie. It looks to me like the article history shows there's been consensus for that list for a majority of the article's existence, ever since "Petekistler" added it in May 26, 2020. Nortis, for example, was adding artists to that list at least as recently as February 2021.https://en.wikipedia.org/w/index.php?title=Artificial_intelligence_art&type=revision&diff=1008351344&oldid=1008343559 I am not really a fan of lists, but that list was in this article when I discovered it back in July, so I did my best to improve the article with it since that seemed to be the desire of previous editors. It seems pretty standard to talk about representative artists in an article about a particular type of art, including in bulleted list format as in for example Impressionism. Is your issue, Ohnotisjamie, that early AI artists like Harold Cohen were in a bulleted list? If so, would you mind if I added Harold Cohen back into to this article, but inside of a paragraph instead of a bulleted list, so this article doesn't read like AI art started in 2014? This article, after its recent changes by Nortis and Drmies is written in a way that makes us look like we are unaware of about 50 years of AI art history. --Elspea756 (talk) 05:55, 14 January 2022 (UTC)
I am going to summarize some of the discussion above. 1) I say we should have examples of contemporary AI art projects as well as "AI art projects from the 1960s through 1990s"." 2) MrOllie agrees, saying "Any art produced by AI, even in the 1960s, would be on topic." 3) CharlesGillingham also agrees, saying "If someone used a technique to create art that came from AI's huge arsenal, in any decade, then we would have to label it as "AI art". Even if they used AI techniques invented in the 50s". 4) Nortis, you have repeatedly attempted to remove all AI art history prior to 2014. Can you please explain your reasoning behind this? 5) After I undid Nortis's most recent attempt to remove these 50 to 60 years of AI art history, Drmies has arrived to apparently support Nortis's version of history. Drmies, could you please explain your reasoning for removing all AI art history prior to 2014? Thank you. --Elspea756 (talk) 04:05, 14 January 2022 (UTC)
When I created the page, I added a few examples of people who where working in this field. It was so the general public could have a better understanding of the people involved at that time as the field was very new. Lets make it clear the people I added, I had no connection to whatsoever, I had spoken to them or ever been in contact with them. It was purely for public guidance. Now there is literary hundreds if not thousands of people working with AI art. There is no need to start listing every single person in this field now, to do so would be ridiculous, no other art subject on wikipedia does this. There are now many sites dedicated to AI artists which list people in this field with full profiles and contact details etc. Clearly it would seem user Elspea756 has an agenda in to include some people on this page for whatever reasons that may be. --Norttis (talk) 07.20, 14 January 2022 (UTC)
- Norttis, we have been over this repeatedly. AI art has existed for over 50 years. Multiple editors who study in this field have all told you this over and over. We have provided multiple sources to try to help you learn about this topic. There is no way you were editing this Wikipedia article when "the field was very new" unless you were somehow editing this Wikipedia article 50 years ago. --Elspea756 (talk) 07:53, 14 January 2022 (UTC)
- Elspea756, who is "we"? I know I can speak of a "we", as in "the people who keep reverting your edits because we don't believe they are helpful" or something like that--but who is your "we"? Because you are the only one bringing that content in. And I have to say that I'm a bit tired of things like "Drmies, could you please explain your reasoning for removing all AI art history prior to 2014?" Well, have you stopped beating your wife? Drmies (talk) 15:28, 14 January 2022 (UTC)
- "We" who have have explained the early history of AI art include myself, MrOllie, and CharlesGillingham. The only two people I see who support the idea that we should remove any description of AI art prior to 2014 are Drmies and Norttis. --Elspea756 (talk) 15:47, 14 January 2022 (UTC)
- Oh stop that bullshit. My reverting your edits does not mean I endorse this or that definition. Drmies (talk) 18:14, 14 January 2022 (UTC)
- Well, other than the profanity, that's great to hear! It sounds like three of us (MrOllie, CharlesGillingham and I) all agree that AI existed before 2014, Norttis is the sole person who disagrees and believes "the field was very new" in 2020, and Drmies holds no opinion? That sounds like we again have strong agreement to include mid-20th-century AI art like Howard Cohen's. Thank you for clarifying, Drmies. --Elspea756 (talk) 19:10, 14 January 2022 (UTC)
- Oh stop that bullshit. My reverting your edits does not mean I endorse this or that definition. Drmies (talk) 18:14, 14 January 2022 (UTC)
- "We" who have have explained the early history of AI art include myself, MrOllie, and CharlesGillingham. The only two people I see who support the idea that we should remove any description of AI art prior to 2014 are Drmies and Norttis. --Elspea756 (talk) 15:47, 14 January 2022 (UTC)
- Elspea756, who is "we"? I know I can speak of a "we", as in "the people who keep reverting your edits because we don't believe they are helpful" or something like that--but who is your "we"? Because you are the only one bringing that content in. And I have to say that I'm a bit tired of things like "Drmies, could you please explain your reasoning for removing all AI art history prior to 2014?" Well, have you stopped beating your wife? Drmies (talk) 15:28, 14 January 2022 (UTC)
I continue to think that efforts to define 'Artificial intelligence art' as a recent phenomenon are misguided. That said, I don't support keeping a laundry list of artists in the article. - MrOllie (talk) 15:32, 14 January 2022 (UTC) To that end I restored the prose mention of Cohen's AARON, with a new third party source that calls it an AI. - MrOllie (talk) 15:59, 14 January 2022 (UTC)
- Thank you, MrOllie. Yes, you and I are in agreement that we should definitely describe AI prior to 2014. I really appreciate your addition of Harold Cohen's important work to this article again! --Elspea756 (talk) 16:03, 14 January 2022 (UTC)
Citation needed for Drmies and Nortis's new definition of AI art in the intro
The second sentence of this article has for over a year read "[Artificial intelligence Art] includes works created autonomously by AI systems and works that are a collaboration between a human and AI system." (Bolding is mine for emphasis.) Drmies and Nortis have recently redefined AI art much more narrowly to say the exact opposite, writing that "The AI art must be created autonomously by a AI system without the collaboration of a human (although a human may help contribute to the input of data, the artwork must be completed entirely by the AI on its own merits)." (Again, bolding is mine for emphasis.) This new definition is not consistent with my reading of the literature on AI art, which is almost always by definition a computer/human collaboration of some sort. For an example from the popular press, The New York Times has described AI artists "using machine learning to create ghostly digital avatars that dance along with live performers."[1] That is very much a machine/human collaboration, and doesn't fit at at all with your new definition that says AI art "must be completed entirely by the AI." So, Drmies and Nortis, can you please provide your sources for this new definition of AI art that you've written into the intro? Thank you. --Elspea756 (talk) 06:32, 14 January 2022 (UTC)
- Elspea756, I have not written that. In addition, you said on my talk page that I was "disregarding the discussion of previous editors", which is of course untrue: there is no discussion here or any other place I know of that I would have somehow disregarded. Au contraire--as I said in my edit summary, it is you who disregard the comments and actions of previous editors (User:Norttis and User:MrOllie), and continue to edit war, against consensus. So I would like to ask you to kindly represent the editorial actions of other editors correctly, and to be more careful in your edits and your comments. Three times now you have reinstated the list of people; it's really time to stop. Drmies (talk) 15:26, 14 January 2022 (UTC)
- Drmies, you in this edit [2] and also and this edit [3] changed this article to say that "The AI art must be created autonomously by a AI system without the collaboration of a human." Do you have a source for this? Also, which specific "comments and actions" of MrOllie do you think I have disregarded? You can see above that MrOllie agrees with me that "Any art produced by AI, even in the 1960s, would be on topic." MrOllie does not seem to agree with Norttis and you that "what ever happened in the 1960's is not relevant on this page." --Elspea756 (talk) 15:38, 14 January 2022 (UTC)
- I can't speak for Drmies's intentions, but the diffs you included also included the arbitrary list of names and promotional sourcing; the more restrictive definition was originally introduced by Nortis. I removed the restrictive clause, though to be fair the previous version which explicitly stated a more expansive definition was not sourced either. There may be multiple definitions of the scope of what's included in "artificial intelligence art," but in any case we need a good source for what that is. My gut feeling is that the title of the article implies a more narrow scope. OhNoitsJamie Talk 16:09, 14 January 2022 (UTC)
- THANK YOU, Ohnoitsjamie! I really appreciate your removing that factually incorrect information that Norttis and Drmies have been insisting on including in the introduction! I did provide a source in my comment above where The New York Times describes AI art as including human/computer collaboration. Would you mind if I added human/computer collaboration to this article, using the New York Times as a source? Thank you again. --Elspea756 (talk) 16:19, 14 January 2022 (UTC)
- Oh stop that bullshit. My reverting your edits does not mean I endorse this or that definition. Drmies (talk) 18:15, 14 January 2022 (UTC)
- Drmies, I am not sure what you are saying here. Are you saying you do not have an opinion on whether AI art can be created collaboratively between a human and AI, vs. AI art must be created "without the collaboration of a human"? I was assuming since you put "without the collaboration of a human" in the article twice, then you were endorsing that statement, but if you are saying you don't have an opinion on this, that's helpful to know. Sorry if I have misunderstood. Thank you! --Elspea756 (talk) 18:35, 14 January 2022 (UTC)
- Assuming you're talking about the paragraphy beginning with "Mr. Skybetter is similarly...", that paragraph is specifically talking about "choreographic software"; the bit about human input being needed was also noted in the narrow definition that Nortis proposed. So no, I don't think that source supports a broad scope for this particular article. If we're talking about using AI as a tool in creating art (versus art created singularly by an AI), perhaps that's better covered in Applications_of_artificial_intelligence#Art. OhNoitsJamie Talk 16:31, 14 January 2022 (UTC)
- That New York Times article describes several examples of human/AI collaboration. The article describes "choreographer Trisha Brown ... employed a program that responded to her dancers’ movements by drawing graphics that were then projected onstage." That's one example of a human dancer collaborating with an AI program. The article also describes "Pontus Lidberg ... set out to use A.I. as a more integral part of his choreography, in rehearsal and performance. The goal: To create a dance that articulated the tension between man and machine, by putting the two together onstage." That's another human/computer collaboration. And, yes, the paragraph you point to says "Mr. Skybetter is similarly adamant that “none of these technologies can really exist without a human hand.” A human is needed at every stage ..." To me, this is so commonsense that it shouldn't need a source (of course humans collaborate with AI to create art! Can you provide sources for AI art where humans do no collaborate with the AI?) but if you are demanding a source for this, is there any agreement among editors here that this New York Times article a good enough source to describe the existence of human artists collaborating with AI? --Elspea756 (talk) 16:54, 14 January 2022 (UTC)
- That article is specifically talking about the application of AI to dance choreography; the article does not attempt to define AI art broadly. OhNoitsJamie Talk 17:57, 14 January 2022 (UTC)
- Ok, so we are in agreement that dance is one art form where artists collaborate with AI, but you would like examples of other art forms where artists collaborate with AI? --Elspea756 (talk) 18:00, 14 January 2022 (UTC)
- We are in agreement that people can collaborate with AI in all sorts of ways. We are not in agreement as to whether the scope of this article broadly includes any AI collaboration/usage. OhNoitsJamie Talk 18:41, 14 January 2022 (UTC)
- I just want to note that, if I now understand correctly, Drmies clarifies up above that they don't hold an opinion on this either way, as to whether AI art can or cannot be a collaboration between humans and AI. So, if I'm understanding correctly, the only person who disagrees completely with the idea that AI and humans can collaborate is Norttis, Ohnoitsjamie agrees artists can collaborate with AI but that might be outside the scope of an article on AI art, and I think the idea that human artists and AI can collaborate is just commonsense and belongs in an article on AI art. It was in the article for over a year, and I have provided a New York Times source for it, and I am open to providing more sources if desired. Are there any other opinions on this that I may have missed? --Elspea756 (talk) 18:49, 14 January 2022 (UTC)
- We are in agreement that people can collaborate with AI in all sorts of ways. We are not in agreement as to whether the scope of this article broadly includes any AI collaboration/usage. OhNoitsJamie Talk 18:41, 14 January 2022 (UTC)
- Ok, so we are in agreement that dance is one art form where artists collaborate with AI, but you would like examples of other art forms where artists collaborate with AI? --Elspea756 (talk) 18:00, 14 January 2022 (UTC)
- That article is specifically talking about the application of AI to dance choreography; the article does not attempt to define AI art broadly. OhNoitsJamie Talk 17:57, 14 January 2022 (UTC)
- That New York Times article describes several examples of human/AI collaboration. The article describes "choreographer Trisha Brown ... employed a program that responded to her dancers’ movements by drawing graphics that were then projected onstage." That's one example of a human dancer collaborating with an AI program. The article also describes "Pontus Lidberg ... set out to use A.I. as a more integral part of his choreography, in rehearsal and performance. The goal: To create a dance that articulated the tension between man and machine, by putting the two together onstage." That's another human/computer collaboration. And, yes, the paragraph you point to says "Mr. Skybetter is similarly adamant that “none of these technologies can really exist without a human hand.” A human is needed at every stage ..." To me, this is so commonsense that it shouldn't need a source (of course humans collaborate with AI to create art! Can you provide sources for AI art where humans do no collaborate with the AI?) but if you are demanding a source for this, is there any agreement among editors here that this New York Times article a good enough source to describe the existence of human artists collaborating with AI? --Elspea756 (talk) 16:54, 14 January 2022 (UTC)
- Oh stop that bullshit. My reverting your edits does not mean I endorse this or that definition. Drmies (talk) 18:15, 14 January 2022 (UTC)
- THANK YOU, Ohnoitsjamie! I really appreciate your removing that factually incorrect information that Norttis and Drmies have been insisting on including in the introduction! I did provide a source in my comment above where The New York Times describes AI art as including human/computer collaboration. Would you mind if I added human/computer collaboration to this article, using the New York Times as a source? Thank you again. --Elspea756 (talk) 16:19, 14 January 2022 (UTC)
- I can't speak for Drmies's intentions, but the diffs you included also included the arbitrary list of names and promotional sourcing; the more restrictive definition was originally introduced by Nortis. I removed the restrictive clause, though to be fair the previous version which explicitly stated a more expansive definition was not sourced either. There may be multiple definitions of the scope of what's included in "artificial intelligence art," but in any case we need a good source for what that is. My gut feeling is that the title of the article implies a more narrow scope. OhNoitsJamie Talk 16:09, 14 January 2022 (UTC)
- Drmies, you in this edit [2] and also and this edit [3] changed this article to say that "The AI art must be created autonomously by a AI system without the collaboration of a human." Do you have a source for this? Also, which specific "comments and actions" of MrOllie do you think I have disregarded? You can see above that MrOllie agrees with me that "Any art produced by AI, even in the 1960s, would be on topic." MrOllie does not seem to agree with Norttis and you that "what ever happened in the 1960's is not relevant on this page." --Elspea756 (talk) 15:38, 14 January 2022 (UTC)
End of article mention of beeple NFT is not AI related
End of article mention of beeple NFT is not AI related
Should be removed or demonstrated to be AI related. 135.23.202.160 (talk) 13:56, 31 May 2022 (UTC)
- I agree, NFTs are an investment bubble-driven fad. I removed it. MrOllie (talk) 14:05, 31 May 2022 (UTC)
Correcting unsourced, false description of a narrow range of AI arts programming
There has been a recent introduction of an unsourced, false sentence saying "There are many other AI art generation programs, ranging in complexity from simple consumer-facing mobile apps to Jupyter notebooks ..." This is an unsourced and false description that AI programs only exist in this very narrow range, since there is of course a long history of AI artists and researchers who are always creating and using new programs that are not available to the general public. There are many types of AI that are simpler than a mobile phone app, and many types of AI that are more complex than a Jupyter notebook. I previously edited this sentence to instead say "There are many other AI art generation programs, including ..." so as not to give the false impression that there are no programs outside of this very limited range. This change was undone by editor Camdoodlebop. The edit summary they used was "reverted vandalism." I am about to remove this false, unsourced "ranging" description again, and I am leaving this talk page message in case anyone wants to discuss whether they have a source saying AI only exists in this very narrow range or whether my correcting this false statement is "vandalism." Elspea756 (talk) 13:15, 1 September 2022 (UTC)
Adding a Stable Diffusion example
Would it be acceptable to add an example from Stable Diffusion that a user created themselves, or would they need to source the example from a third party to avoid self-promotion? I have some examples I've uploaded to Wikimedia I could add. Camdoodlebop (talk) 00:41, 10 September 2022 (UTC)