4 ms·

DALL-E's docs for example mention it can output whole copyrighted logos and characters[1] and understands it's possible to generate human faces that are bear th

DALL-E's docs for example mention it can output whole copyrighted logos and characters[1] and understands it's possible to generate human faces that are bear the likeness of those in the training data. We've also seen people recently critique Stable Diffusion's output for attempting to recreate artists' signatures that came from the commercial trained data. That said by a certain point the kinks will be ironed out and likely skirt around such issues by only incorporating/manipulating just enough to be considered fair use and creative transformation. [1] "The model can generate known entities including trademarked logos and copyrighted characters." https://github.com/openai/dalle-2-preview/blob/main/system-card.md https://github.com/openai/dalle-2-preview/blob/main/system-c...

TulliusCicero 4y ago
Right, but literal logos or obvious copyrighted characters aren't the same thing as a general style. I don't believe style is copyrightable. If there's a source that says otherwise, I'd love to read about it.
- Springtime 4y ago
  What I was meaning is that the models have the potential to output wholesale sections of original trained data, not merely interpreting a style. Which ties into grandparent's point about how in music if a portion of music is sampled by another artist for a separate piece of music they are obligated to pay royalties for it. As I said though those things will likely be ironed out to avoid this more in time.
- wnkrshm 4y ago
  It's not illegal, but if an artist does it, like deliberately copy another artist's style, as in more than a tribute but an attempt at carbon copy, that artist is almost shunned. So the nonlinear diffusion systems out there are told by prompt to make what in the art world is considered to be very rude.

4 ms·

DALL-E's docs for example mention it can output whole copyrighted logos and characters[1] and understands it's possible to generate human faces that are bear th

by Springtime 4y ago

TulliusCicero 4y ago
Right, but literal logos or obvious copyrighted characters aren't the same thing as a general style. I don't believe style is copyrightable. If there's a source that says otherwise, I'd love to read about it.
- Springtime 4y ago
  What I was meaning is that the models have the potential to output wholesale sections of original trained data, not merely interpreting a style. Which ties into grandparent's point about how in music if a portion of music is sampled by another artist for a separate piece of music they are obligated to pay royalties for it. As I said though those things will likely be ironed out to avoid this more in time.
- wnkrshm 4y ago
  It's not illegal, but if an artist does it, like deliberately copy another artist's style, as in more than a tribute but an attempt at carbon copy, that artist is almost shunned. So the nonlinear diffusion systems out there are told by prompt to make what in the art world is considered to be very rude.