dear nettimers, do you have any ideas/comments on the copyright case against 
stable.diffusion?
 https://www.eff.org/deeplinks/2023/04/how-we-think-about-copyright-and-ai-art-0

In particular, what triggers my email is that a proprietary stock image company 
has sued stable diffusion and (afaik) not the "proprietary"ai or the other 
popular proprietary ai text-to-image service. any ideas why? just because they 
had proof that their images were used for training stable diffusion? i do not 
think so… I suspect that it might be a proprietary vs floss models issue. 
stable diffusion's license is (again, afaik) not an fsf or osi-approved floss 
license but at least it is a kind of "permissive" license. 

by the way, i find the discussion about whether the output of deep learning 
text-to-image models would be the copyright of the users, the developers, or 
the copyright holders' of the data used for training the models, simply 
ridiculous. it is simply a.powerful demonstration of the fact that culture 
builds on the past and to reciprocate, to encourage building on your works 
freely makes the culture richer, and more enjoyable by all. what makes 
these.models possible is that many people have created images/works in the past 
amd they are made somehow "accessible" to the world on the internet, with 
metadata, which is mostly created by many others. the combination of all these 
is the collective intelligence/heritage of humanity. the outcome of what became 
possible mainly because of the people of the world would belong to the people 
of the world, thus should be in the public domain, as in the output of stable 
diffusion.that's very clear to me. some may even argue that they might be 
copylefted to prevent their commercial exploitation. but in anyways, they 
shouldn't be the property of any individual,.or company. 
--
özgür k.
gpg:A3E6 57AD E14D 1F66 A546 6101 BA42 0724 E750 C5AE
-- 
# distributed via <nettime>: no commercial use without permission
# <nettime> is a moderated mailing list for net criticism,
# collaborative text filtering and cultural politics of the nets
# more info: https://www.nettime.org
# contact: [email protected]

Reply via email to