Hacker News Reader

Comment by TheMiddleMan | Hacker News Reader

An artist can look at images for reference, and draw something new inspired by them. Why does it matter if a software tool can do this much faster?

If the artist makes the image very similar to one of the reference photos, it may be a copyright violation. It doesn't matter if the artist used a pencil or software to create the new work.

Current AI image generation does, however, make it easy to unknowingly violate copyright. If it generates an image similar to something else out there you wouldn't know.

I don't know much about copyright law though, am I wrong?

View on HN

Well, there was a copyright case in Europe recently where an artist had taken a photograph, flipped it horizontally, and painted it.

It was deemed an original work by the court.

I can’t see how, with such a precedent, they could rule that SD doesn’t produce original works.

https://www.rangefinderonline.com/news-features/industry-new...

> It was deemed an original work by the court.

The resolution is much weirder than that, the court argued that the pose isn't original enough for the photo to deserve copyrights at all, independently of what the plagiarist did with it.

Iirc it can be a problem copyright wise if I paint from a photo reference because it can infringe the photographers IP.

I dont think anyone can be sued for making a drawing of some photo.

You can claim intellectual property for comic characters etc, but not photos.

YOu could get someone to go get a similar photo taken at the same place, and use tools to enhance it

There is a somewhat popular lawsuit right now which argues exactly that (and is reported to go into the next instance): https://petapixel.com/2022/12/08/photographer-loses-plagaris...

I don't know if you're right or wrong, but it seems plausible that we could create a database of copyrighted images to check against.

Only the training set should suffice.

Every original image is copyrighted. You're suggesting making a digital copy of every image there is to check that AI isn't generating digital copies of every image there is.

Not a copy, a hash or fingerprint. Just enough data to measure if it's substantially similar.

But yes, it may be infeasible to index and compare against every image ever uploaded.

If I understand correctly, wouldn't a hash database of <just the training set> be larger than the actual model? (in fact by 1 or 2 orders of magnitude?)

Approximately, yes.

Yeah, I guess so. The models are only 4 or 8 GB. A giant list of hashes would be bigger, sure. But they're 2 very different things. Model is for generating new images, this hash database is copyright enforcement. If you really want to check for violations I don't know how else you're going to do it.

Couldn't I just add a few non-sense bytes into my images to change the hash/fingerprint?

Hash yes, fingerprint maybe no. Maybe I'm using the term incorrectly here, but I think of fingerprint like a lossy hash. Like one way of doing this would be to resize the image to, say, 8 by 8, and quantize it to say, 16 colors. So the fingerprint size is 884 bits=32 bytes. Tiny changes aren't likely to change the fingerprint. You'd probably have to do something a little more clever so as not to get too many false positives though. Or once you get a hit, do a deeper comparison.

https://www.wired.com/2011/01/hope-image-flap/amp

Reply on news.ycombinator.com