Stable Attribution

This appears to be just looking for the nearest neighbors of the image in embedding space and calling those the source data. This by definition would find similar looking images, but it's not strictly correct to call it attribution. To some extent all of the training data is responsible for the result - as an example, the model is also learning from negative examples. The result here may feel satisfying, but it's overly simplistic and it's misrepresenting what it is to call it attribution.

(The silly story they have on the site doesn't really score any points either, it reminds me of RIAA et al)

I just tried this with an image I took with my phone and it gave me 10-15 images that the "ai" used to generate my image, proving this is an absolute fraud of a concept.

I assumed fraud, but also just dumb.

The genie isn’t going back on the bottle.

This is a great website, but not in the way the authors intended. Based on some of the examples they explicitly provided, it is clear to me Stable Diffusion creates novel art. Here's a random example https://www.stableattribution.com/?image=a2666aee-0a1a-411b-...

I will admit this is a nice tool for verifying the creations of SD aren't pure copies, so I think it will be useful for a time. But as AI-generated images start to taint future datasets, attribution is going to be significantly more complicated.

Am I the only person who looks at this and thinks, “So what?”

I mean it’s technically impressive, I guess. But why would anyone care or pay for this product?