The shoes part is the most plausible reason for this image being, if not AI, at least heavily doctored. 5 pairs of shoes, all the same style (canvas with white rubber sole), three of them with the rear foot at the exact same angle with the exact same bend... The odds of that happening in real life are astronomically low. Also look at the odd posture of the person in the middle of the rear row - no one walks like that on a flat surface, and Their front foot is the only one of the group with a super dark "shadow" under it.

It was on the Is This AI Reddit sub. Also, I think the AI has put a hinge under the lock (down by the yellow sign).

Uhm, looks like my comment was unclear. I don't know if this is AI generated or not, it might be. In any case, there wouldn't be a source for or against that.

What I'm referring to is the claim that it saves energy to generate identical features. My understanding was that with stable diffusion, the energy usage would be proportional to the number of pixels and number of iterations, not what features end up on the picture.