Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: DALL-E was trained on watermarked stock images?

news.ycombinator.com

Post reply on HN
Print view View original

161–170 of 233 posts

Re: Ask HN: DALL-E was trained on watermarked stock images?

#161

Post by mminer237 »

anigbrowl wrote: post #57
snickerbockers wrote: post #50

Earlier quoted context omitted.

That's technically not a stock image, it's a portait that has been public domain for a long time.

But you've seen many PD images reshared by stock imagery companies. It raises the question of why false assertions of ownership aren't easily prosecuted, given that they constitute a kind of fraud upon the public.

That's what trademark law is. But I don't think the person who painted the portrait is going to be suing Getty for impugning his good reputation.

If something's public domain, anybody can use for anything they want, even if that's just rehosting it with your watermark.

Re: Ask HN: DALL-E was trained on watermarked stock images?

#162

Post by antioppressor »

chrismorgan wrote: post #87
_trampeltier wrote: post #82

If you read the licence from Getty, they say, you are not allowed to use Getty pictures for ML.

What that license text says is irrelevant, because they’re not using it under that license, but under fair use exemptions in copyright law.

Fair use, until it disrupts market, sidelines creators, gobbles up the market, make it ubiquitous, lock people in, then let the regulators craft some bs law that will change nothing, compensates noone. ;)

Re: Ask HN: DALL-E was trained on watermarked stock images?

#163

Post by omnimus »

gojomo wrote: post #90
saurik wrote: post #74

Earlier quoted context omitted.

I'm still not seeing the "transformative" argument: the point of transformation isn't "it is in a different format" but (to quote Wikipedia, which is, of course, dumb... I'm sorry ;P) where one "builds on a copyrighted work in a different manner or for a different purpose from the original". The reason a search engine thumbnail is transformative isn't because it has been transformed to make it smaller... it is becaus…

Autogenerated, often fantastical, never-seen-before AI images strike me as a paradigmatically 'transformative' use. It's novel. It's shocking to many practicioners how flexible & high-quality the images can be. It will unlock all sorts of new downstream creation. The representation that feeds the generation is statistical, even to the point of being plausibly factual : these things/people/places/concepts can be abstr…

"AI artist" doesn't add any of its "own flair". It builds exclusively on past experience and work of humans. And it also directly completes with them without any thought of credit or compensation.

People are really underplaying how damaging this is going to be for the industry. It's going to completely decimate it. You can already see people using names of artists in the DALL-E prompt to get "their" work for few dollars avoiding any copyright or social issues.

Artists will suddenly be competing with AI on price and time - why we should pay you living wage when we instantly generate something close enough.

Why would anyone try to create some new aesthetic or push anything further if their effort will be replicated next week when the model gets updated with new source data. Everything is gonna get stuck to aesthetic of 2025 and before.

It's completely inhuman.

Re: Ask HN: DALL-E was trained on watermarked stock images?

#164

Post by yreg »

BeefWellington wrote: post #157
yreg wrote: post #102

Earlier quoted context omitted.

Unlike Copilot, DALL-E et al. don't produce verbatim copies of trained data. Copying ideas and styles has always been a fundamental part of art history, so an artwork right holder might have a hard time successfuly sueing a user for the user's generated image looking similar to the right holder's artwork.

"Verbatim" is an interesting term since I'm not certain it matters. In this case OP here demonstrated DALL-E generating a trademarked watermark on top of an image. I doubt the courts, looking at that, would believe that that's not close enough to their trademark to infringe. The art world's copyright suits are all over the place in terms of what's sufficient to meet the threshold of "fair use" or "not a copy". It's h…

The case you refer to (Cariou v. Prince) is also a case where part of the artwork is reproduced verbatim :)

Re: Ask HN: DALL-E was trained on watermarked stock images?

#165

Post by omnimus »

int_19h wrote: post #108
egypturnash wrote: post #37

Earlier quoted context omitted.

My feeling is that while these things may be technically falling under fair use, I really feel like they are running roughshod over a lot of ethical and moral lines and that perhaps "fair use" needs to be redefined to explicitly exclude this kind of processing. And if it kills these things, oh well. "Being an artist" is a precarious enough existence in this world as is, I'd be delighted to stop worrying about having…

You will still be competing against it. It'll just be outsourced and presented as manual work.

Exactly this. It will just become tool used by those who can afford it in secret. No matter what copyright it breaks.

Re: Ask HN: DALL-E was trained on watermarked stock images?

#166

Post by drran »

colejohnson66 wrote: post #24
alcolade wrote: post #15

Earlier quoted context omitted.

Another human analogy could be: you take a photo from every art show and museum, and use those for reference as you paint.

The analogy of using your cortex is more apt. My understanding of neural networks is that there are no remnants of the original inside it. The training data is used to back propagate a bunch of weights. Your brain works like those neural network neurons; they learn when to fire, but they don’t know the intricate detail like a photo. Hence why many claim eyewitness testimony is bogus.

> My understanding of neural networks is that there are no remnants of the original inside it.

Can you prove that? I can prove the opposite.

Re: Ask HN: DALL-E was trained on watermarked stock images?

#167

Post by JacobiX »

supermatt wrote: post #135
JacobiX wrote: post #128

The first thing that I try after generating an image from DALL-E is using reverse image search. I do it on every image that I intend to use, more often than not, I find a very similar image, in this case I discard it and vary my prompts.

> more often than not, I find a very similar image Can you give an example? I was also doing reverse image searches, and I havent seen a single case of an image being closely related to another unless it was used as the base for inpainting.

At first, you can experiment with "extreme" input that can cause overwriting like: a photo of Marylin Monroe, etc. To search using reverse image search I use yandex, and I downsample the image.

Re: Ask HN: DALL-E was trained on watermarked stock images?

#168

Post by drran »

ale42 wrote: post #119
donkarma wrote: post #22

Earlier quoted context omitted.

yeah except this artist wont go around painting watermarks

There's a big difference here. DALL-E (or any other similar tool) is not an artist, it is a commercial product that can produce something that might be considered art. The usage might be ok, but the product itself might not... if I paint using stolen pigments (not saying that DALL-E is alike to this), my painting would be legal, but the fact of having stolen the pigment is very illegal...

Product of machine cannot be copyrighted. Only humans have the right to copyright something. Dall-e like kaleidoscope then. Anybody, even monkey, can generate similarly looking images using Dall-e with the same prompt. So there is two options:

a) Dall-e images are without copyright.

b) Original authors of work used by Dall-e can claim copyright over images generated.

Post reply on HN