Engadget (from March 16):

Any images that are produced by giving a text prompt to current generative AI models, such as Midjourney or Stable Diffusion, cannot be copyrighted in the US. That's according to the US Copyright Office (USCO), which has equated such prompts to a buyer giving directions to a commissioned artist. "They identify what the prompter wishes to have depicted, but the machine determines how those instructions are implemented in its output," the USCO wrote in new guidance [pdf] it published to the Federal Register.

"When an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the 'traditional elements of authorship' are determined and executed by the technology — not the human user," the office stated.

[...]


"Solely" is probably the essential word here. And it strikes me as a sensible ruling - we don't want some server farm generating exabytes of images just to serve as fodder for copyright lawsuits.

"Your picture of a sunset is clearly a copy of our copyrighted image X3g2wCBBa271-$$.jpg Pay us for a license or we'll sue you into poverty. Have a nice day."

:-/

If you take what the AI generates and edit it in any creative way, then it's your work and should be protected, I would think. But check the PDF.

Good luck!

Cheers,
Scott.