Disney has assembled a very good case against one generative AI company, showing that the stuff “learned” from shared art is essentially infringing on copyright and failing to assemble its own ideas. They asked it for “A man in a metal suit” and it would consistently give them Iron Man, down to the fine details.
We’ve moved away from the simple Copy and Paste tools in image editors, but I would say not nearly far enough for technology to evade the principles of trademark and copyright infringements. The “model” is in many ways just a database of other people’s stolen artwork, ready to be rearranged.
Even so, if you commission a picture of Iron Man or if you generate a picture of Iron Man… both are going to be covered under trademark without any need to differentiate the two.
Disney has assembled a very good case against one generative AI company, showing that the stuff “learned” from shared art is essentially infringing on copyright and failing to assemble its own ideas. They asked it for “A man in a metal suit” and it would consistently give them Iron Man, down to the fine details.
We’ve moved away from the simple Copy and Paste tools in image editors, but I would say not nearly far enough for technology to evade the principles of trademark and copyright infringements. The “model” is in many ways just a database of other people’s stolen artwork, ready to be rearranged.
Even so, if you commission a picture of Iron Man or if you generate a picture of Iron Man… both are going to be covered under trademark without any need to differentiate the two.