The linked post on their website has a few more sentences, but no hard promises yet…
Thanks to our diligent community who pointed out a series of recent marketing images may have included elements of generative AI, we are rethinking our process of how we work with vendors for our marketing creative.
We already made clear that we require artists, writers, and creatives contributing to the Magic TCG to refrain from using AI generative tools to create final Magic products. What’s now apparent is that we need to update the way we work with vendors on creative beyond our products—like marketing images we use on social media—to make sure that we’re supporting the amazing human ingenuity that is so important to Magic. Along with so many others, we also want to get better at understanding whether and how AI is used in the creative process. We believe everyone benefits from more transparency and better disclosure. We can’t promise to be perfect in such a fast-evolving space, especially with generative AI becoming standard in tools such as Photoshop, but our aim is to always come down on the side of human made art and artists.
I had a look at the marketing image they used, it appears to be a fully generated image (with the card faces inserted afterwards). It isn’t really anything to do with the plugins in Photoshop. That’s a deliberate conflation of things like style transfer or inpainting with full generation via diffusion from random noise. The former starts with an image input which is presumably not made by AI
I think Photoshop does actually have a graphics art ai market plugin, where AI artists can (allegedly) make a bit of money when their stuff is picked out of the catalog. I think the idea is you can grab something like stock assets to mix into your work, but there’s so much of it that it might as well be unique
This still sounds like an excuse/non-apology, but it’d be feasible if I thought their company had a moral center at this point
Yeah I’m saying the image in question specifically looks like it was generated from random noise in Stable Diffusion though. Just based on my experience using it, there’s a few telltale errors (aside from what’s been highlighted by others) which would commonly be unnoticed/ignored by an intermediate user. Things you can’t unsee once you learn to see them, like how the model was confused about where the windowsill ends (or whether it’s a windowsill or a table beside the window)
Oh yeah, for sure - it also has that trademark out of focus background the newer models tend to use, and the fact that it looks more crisp than it should at that pixel count
I don’t believe their excuse, but this could have been in that marketplace under backgrounds… Once I looked at it, it seems like the artist just took the card faces and warped them to line up with the perspective of the scene and called it a day
The linked post on their website has a few more sentences, but no hard promises yet…
Less than a month ago
Lol
I had a look at the marketing image they used, it appears to be a fully generated image (with the card faces inserted afterwards). It isn’t really anything to do with the plugins in Photoshop. That’s a deliberate conflation of things like style transfer or inpainting with full generation via diffusion from random noise. The former starts with an image input which is presumably not made by AI
I think Photoshop does actually have a graphics art ai market plugin, where AI artists can (allegedly) make a bit of money when their stuff is picked out of the catalog. I think the idea is you can grab something like stock assets to mix into your work, but there’s so much of it that it might as well be unique
This still sounds like an excuse/non-apology, but it’d be feasible if I thought their company had a moral center at this point
Yeah I’m saying the image in question specifically looks like it was generated from random noise in Stable Diffusion though. Just based on my experience using it, there’s a few telltale errors (aside from what’s been highlighted by others) which would commonly be unnoticed/ignored by an intermediate user. Things you can’t unsee once you learn to see them, like how the model was confused about where the windowsill ends (or whether it’s a windowsill or a table beside the window)
Oh yeah, for sure - it also has that trademark out of focus background the newer models tend to use, and the fact that it looks more crisp than it should at that pixel count
I don’t believe their excuse, but this could have been in that marketplace under backgrounds… Once I looked at it, it seems like the artist just took the card faces and warped them to line up with the perspective of the scene and called it a day
1: type “bokeh”
2: hold Shift + ⬆️
3: ???
4: PROFIT!