crecente,
@crecente@games.ngo avatar

Assume a website plans to use user-contribution content to train LLMs. The license for the content is CC BY-SA.

❓ Would the output from the resulting LLMs be required to provide attribution?

@law #LLM #AI #copyright #creativecommons #technology #license

twan,
@twan@mastodon.online avatar

@crecente @law That would depend on whether the output is considered to be a derivative, I guess? (Assuming it's not a quote.)

1dalm,
@1dalm@deacon.social avatar

@crecente @law

Not likely in the US. I personally believe that it's unlikely any of the copyright claims against AI generators will stand.

If Google Books -a web site that literally copied full copyrighted works and allowed people to search through them and showed them actual images of the search results- was deemed to not be in violation of copyright law, then I just don't see how AI systems are going to be found in violation.

Ultimately, if just analyzing and learning from prior works and using that knowledge to create derivative works is copyright violation, then literally every single human that has ever drawn or written anything at all is also in violation.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • llm
  • rosin
  • Youngstown
  • khanakhh
  • ngwrru68w68
  • slotface
  • ethstaker
  • mdbf
  • everett
  • kavyap
  • DreamBathrooms
  • thenastyranch
  • cisconetworking
  • magazineikmin
  • Durango
  • megavids
  • GTA5RPClips
  • anitta
  • tester
  • tacticalgear
  • InstantRegret
  • normalnudes
  • osvaldo12
  • cubers
  • provamag3
  • modclub
  • Leos
  • JUstTest
  • lostlight
  • All magazines