File talk:Text-to-image woman underwater.png

Which AI was used?

@Camdoodlebop: Which AI was used/directed to make this image? From this edit I suspect it's also Stable Diffusion, is that true?

Please move your images to the appropriate subcategories of https://commons.wikimedia.org/wiki/Category:Artificial_intelligence_art.

Thank you (also thanks for uploading these good images here). Prototyperspective (留言) 22:31, 17 September 2022 (UTC)

@Prototyperspective: Yes, it was Stable Diffusion! I generated that and the other similar images back when it was just on a discord server during the beta testing stage. I tried to upload some examples of how it can create photorealism as I hadn't seen much of that on Wikipedia! Thanks for the advice as well, I'll be sure to get them in the right categories. Do you know how I would do that to existing uploads? Camdoodlebop (留言) 04:58, 18 September 2022 (UTC)

• Click on Preferences at the very top of this page. Then click on the tab "Gadget". There search for "HotCat" and check it (blue checkbox). Then click save. (And all of this shouldn't be necessary as it should be the default, I'll ask about that later elsewhere.)
• Once you have done that open the page of the image and you see a + and a - and ± signs next to the categories at the bottom of the page. Click on ± to rename "Artificial intelligence art" to "Stable Diffusion" (and click Save or Ok).
• You could also add some other categories depending on the style of each artwork, for example "Photorealism" if that's the case(! do check if it is) for the image (I think such categories could and should be added despite it being an AI-generated image as long as there is no flood of such images here – it may even be used to help illustrate articles that currently don't have an image).

Prototyperspective (留言) 09:42, 18 September 2022 (UTC)

i think the picture is public domain because the ai created it indirectly through you so you > stable diff > ai generated picture hence no copyright Slinkyw (留言) 01:00, 18 February 2023 (UTC)

Good point, maybe the licenses should be changed. However, I think that's a misconception...the AI generates the picture but it didn't do so independently, it's used as a tool by giving it (typically well-engineered) prompts and selecting the best picture (or readjusting the prompt).

Nevertheless, I'd support it anyway because it would allow the upload and use of nearly all the images generated by Stable Diffusion which were posted online, many of which e.g. could be useful for illustration of various Wikipedia articles.

Is there a discussion about this somewhere? Prototyperspective (留言) 10:02, 18 February 2023 (UTC)