Talk:Artificial intelligence art

Should we refer to AI art as art? By some definitions it is not, and I think it is important that the article reflects this!
Some define art as being a form of human communication, which would exclude AI art from being art. I think we should reconsider statements such as "Artificial intelligence art is any artwork, particularly images and musical compositions, created through the use of artificial intelligence (AI) programs, such as text-to-image models and musical generators.", as they imply that AI art is art without taking into account the lack of consensus on the topic. Rhinoceros Beetle (talk) 15:43, 9 May 2023 (UTC)


 * Please get acquanted with the definition of Art.
 * Specifically Purpose and Steps
 * https://en.wikipedia.org/wiki/Art#Purpose
 * https://en.wikipedia.org/wiki/Art#Steps
 * I will bring your attention, to the fact that the first two steps of the creation process of art.
 * The Preparation (1st Step) - "In the first step, the artist envisions the art in their mind. By imagining what their art would look like, the artist begins the process of bringing the art into existence. Preparation of art may involve approaching and researching the subject matter."
 * The Creation (2nd Step), states: "In the second step, the artist executes the creation of their work. The creation of a piece can be affected by factors such as the artist's mood, surroundings, and mental state."
 * And of course, please get acquainted with the term and definition Conceptual Art (not to be confused with concept art)
 * https://en.wikipedia.org/wiki/Conceptual_art#Conceptual_art_and_artistic_skill
 * "Conceptual art, also referred to as conceptualism, is art in which the concept(s) or idea(s) involved in the work are prioritized equally to or more than traditional aesthetic, technical, and material concerns."
 * https://en.wikipedia.org/wiki/Art#Skill_and_craft
 * "A common contemporary criticism of some modern art occurs along the lines of objecting to the apparent lack of skill or ability required in the production of the artistic object. In conceptual art, Marcel Duchamp's Fountain is among the first examples of pieces wherein the artist used found objects ("ready-made") and exercised no traditionally recognised set of skills.
 * Tracey Emin's My Bed, or Damien Hirst's The Physical Impossibility of Death in the Mind of Someone Living follow this example and also manipulate the mass media. Emin slept (and engaged in other activities) in her bed before placing the result in a gallery as work of art.
 * Hirst came up with the conceptual design for the artwork but has left most of the eventual creation of many works to employed artisans."
 * The process of creation AI art (or AI assisted art), first of all does not contradict the creation process of Art. And all 3 basic steps can be followed: Envision, Execute and Appreaciate.
 * Also, AI art (or AI assisted art) falls under the Category of Conceptual Art, a modern art form, which you might not like, but it does exist.
 * Thus, using Midjourney and writing your own prompts and evolving your images, when you have a very specific vision in mind does constitute as Art, by Art's own definitions.
 * And just like with any other format, there will be those who are not good at it, who have the imagination of a snail, and there will be those who will create masterpieces of art.
 * Just because, you disagree with existing definitions about what is needed to make an art piece, does not invalidate existing definition of what can be considered art or not.
 * "Art is a diverse range of human activity, and its resulting product, that involves creative or imaginative talent generally expressive of technical proficiency, beauty, emotional power, or conceptual ideas." 82.131.6.207 (talk) 16:48, 22 January 2024 (UTC)
 * On the second of these premises, namely "the artist executes the creation of their work," surely there exists at the very least a level of ambiguity as to whether the artist actually executes that work? The artist in question may have a particular vision in mind when giving a prompt, but this is no guarantor of any form of continuity between what is thought and what is generated, given the mind of the artist is not a source an engine such as Midjourney can access, and could in fact give undesirable results.
 * On top of this, an artist being commissioned to produce an artwork on a set of specified details does not make the commissioner of the art the creator of the art, but rather the artist who produces the final work. Under this context it would make more sense for the person inputing a prompt to be viewed in the role of a commissioner, and the AI model in that of the 'artist', so to speak, but with this the effect had undermines the pre-assumed position that the person created the art, resulting in the generated image being product of machine activity, not human activity. 2A02:C7C:62D9:5C00:9C51:870E:BA29:14D4 (talk) 01:33, 1 March 2024 (UTC)
 * Latent space is merely a new medium. I believe the actual disagreement comes from the fact that technically you could describe it as "algorithmic art" - but that already exists outside genAI; and AI art can encompass more than just works created with diffusion models, which is what most think of when the topic is discussed.
 * The execution piece is interesting because you could just as easily argue there being no guarantee of continuity between an artists mind and canvas with traditional art. Sure, some are able to translate thoughts into the physical realm with relative ease while others are not and I'd argue that spectrum is just as apparent when using these new systems. If one lacks the skills to transfer the idea in their head onto canvas, we would not say they're unable to make art, no? Neural Nets are simply another cog in the "machine" that is taking an idea from the mind's eye and translating it to the physical world. I don't believe anyone has the authority to dictate where that line is, or else we wouldn't consider any digital art as art.
 * To your second point; I am very much under the impression that this view is bred out of a lack of knowledge - I find it very hard to equate AI use to commissioning another human due to the fact that the amount of control one can exude unto a diffusion model is worlds beyond what one is able to do to another human (when commissioning). There is much more to the world of AI art than typing words into a box and clicking "go". I think if one operates under the notion that AI art is "created by a computer" full stop then they would also have to believe in computer sentience.
 * I agree that we need more precise language than "AI art" but, while it's certainly controversial, AI art is art. The end of the third paragraph in this article addresses your point, I feel. Sojoelous (talk) 23:36, 8 May 2024 (UTC)
 * I agree that we need more precise language than "AI art" but, while it's certainly controversial, AI art is art. The end of the third paragraph in this article addresses your point, I feel. Sojoelous (talk) 23:36, 8 May 2024 (UTC)

Slanted towards recent events
There have been some attempts to ignore 20th century AI art and rewrite portions of this article with unsourced claims like "Artists began to create AI art in the early 21st century" There are multiple sources in the article describing AI art before the recent AI boom of the 2020s, so let's not try to rewrite this to only focus on the most recent events. I have added the tag to this page saying "This article appears to be slanted towards recent events. Please try to keep recent events in historical perspective." Thank you. Elspea756 (talk) 00:52, 15 May 2024 (UTC)


 * The entire article needs a rework. It feels incredibly disjointed. Sojoelous (talk) 07:49, 22 May 2024 (UTC)

Lacking Expansive Image Examples
The article seems to include only amateurish or research styled images (astronaut riding a horse on the moon and "cyberpunk" robot). Even taking into account the (at this point old) viral examples, I believe it does not represent the breadth of images out there. Considering it's against the rules and very much frowned upon to upload ones own work - do you have any ideas on how we might expand the article in this regard? To be clear, my reasoning is not "these images stylistically suck" but rather they do not represent even a fraction of what's possible with current advancements in the field.

As an aside, I earnestly think the images of the pope, Trump getting arrested, and the mouse do not belong simply because they are memes and for the purposes of this article not art (or they should at least be smaller than the other examples). Sojoelous (talk) 07:48, 22 May 2024 (UTC)
 * Yes, the current images are almost completely weighted towards user-generated images from amateur artists from the past two years. This has been an ongoing problem. And, yes, we shouldn't have people uploading their own work into this article. We should remove probably half of the current images, particularly the unsourced user-generated images, and replace them with reliably sourced museum-quality artwork from notable professional artists from the entire history of AI art. I will work on this at some point soon. In the meantime, the following unsourced user-generated images could be easily removed: the cow, the landscape, the robot, and the astronaut Elspea756 (talk) 13:30, 22 May 2024 (UTC)
 * Considering the first image appears to be generated by another user (not even considering it's an old Dalle one) - could we just... replace it with a better one? Or at least, a more up-to-date example?
 * Outside of that, I'm not sure what the policy is on using something from say Holly Herndon. Sojoelous (talk) 11:29, 1 June 2024 (UTC)

Edit War
I know I haven’t been too involved with this article, but I keep seeing edits and reverts. This makes me think there is some kind of edit war going on.CycoMa1 (talk) 01:17, 1 June 2024 (UTC)
 * From what I can see, PK-WIKI has added the same image to the article at least three times and it has been removed at least three times by three different editors, The Squirrel Conspiracy, Belbury, and myself . There is a discussion above about the article being slanted toward recent events, and another discussion above about too many examples of meme images. Besides just being another recent meme image that had been already removed by two previous editors, I thought the creation of a new section called "Social and political" with just this example was unnecessary and out of place with the rest of the article, since there are already many examples of social and political use of AI art already described in the article, including Trump memes and Stephanie Dinkins' award for creating AI art based on the interests and cultures of people of color.Elspea756 (talk) 01:38, 1 June 2024 (UTC)
 * There's no "edit war"... I added the image, reverted WP:UCR, then added additional sources in response to an editor questioning the importance of AI to the image.
 * I have not added this image because it is "slanted toward recent events", but rather because the "mega-viral" spread of this AI-generated image made world-wide headlines with major significant coverage of the AI image itself in: The Washington Post; BBC; Al Jazeera; The New York Times; NBC News; Los Angeles Times; etc. That brings it beyond being "just being another recent meme image"; WP:DUE weight requires this article to cover a specific piece of AI art that has been shared by 50 million users, had headlines written about it in every major global newspaper, and (according to reliable sources) for the first time raises major questions about the use of AI art in protests. PK-WIKI (talk) 02:04, 1 June 2024 (UTC)
 * Sorry for the lazy comment.
 * What I mean to say was I kept seeing reverts here.CycoMa1 (talk) 02:44, 1 June 2024 (UTC)

Article on AI art vs. surrealism
https://www.washingtonpost.com/style/of-interest/2024/06/30/ai-art-facebook-slop-artificial-intelligence/ Mapsax (talk) 00:50, 5 July 2024 (UTC)