> and you and i can stand at the same spot and take the same photo with the same camera and it will turn out the same.

But both of our photos would still be protected by copyright despite being near identical to each other. Yet the AI image, created by giving it the exact same factors as input, down to film stock and exposure time, lens selection, time of day and weather conditions etc, would not be.

And there would be some differences in our photos, mostly random factors like the exact pattern of clouds, or visible lights on/off at the moment, passing birds in the sky etc.

And that random factor is something to consider, random imagery generated by nature is copyrightable in an image - like a cloud pattern, vegetation or natural landscape - but not if that random imagery is generated by AI.

> you and i can use the sample computer and provide the same prompts and we will get something different for the same prompt.

The more specific information we give the prompt, the more similar the results would be. I bet we could get pretty close by being providing enough information in the prompt, with the minor differences in detail being reasonably considered inconsequential.