Open Access Sources For Ai Art Data

First of all, please don't get into giant arguments here. I don't want to get in trouble.

I know some of you probably think I hate AI art. I dislike how it has been implemented, not the whole idea.

I happened to look up something today and came across this.

https://www.metmuseum.org/about-the-met/policies-and-documents/open-access#get-started-header
https://www.metmuseum.org/art/collection/search?searchField=All&showOnly=openAccess&sortBy=relevance&pageSize=0

The Met says they have more than 492,000 works published to the public domain, and even have an API to use the library. I did not count them to verify. The Met has been doing this since 2017.

It seems like this would be an ideal source to start with because it's a museum. They don't publish poor quality images, they have designed the site to allow downloading, and it says the use is unrestricted, even for commercial. Other organizations have open access work listed, including the Yale University Art Gallery (looks like image by image) and the Art Institute of Chicago (can sort by open access/public domain). Getty (the museum/research organization) has 81,000+ works from the Museum and 78,000+ archived materials from the Research Institute available for free download. Many of these appear to be growing and new works are being added. There are many more of these collections.

What I'm trying to say, is why did AI researchers not use these first? This would help to eliminate arguments about what can and can't be done with the images. If I was a researcher I'd much rather have museum quality images, most of which have accompanying data, than random pixelated images off the internet. I have ran a few searches from https://haveibeentrained.com/ and it seems to have a ton of duplicates too (the same image is listed over and over from different websites). I know that Laion databases have billions of images listed, but what good does that do if the same image appears twenty or even a hundred times (or more)?

If generative AI companies wanted to improve public access, couldn't they, say, donate to the Met and Getty research to expand the public domain collections (and other organizations)? They could start an open access collection that artists can place their work in too!

It just doesn't make sense to me.