Why *are* the faces always these same 3-5 faces that are all just slight variations of each other? The creator can use the AI to change up the scene and the outfit but the faces can’t be changed significantly? I’m genuinenely curious as someone who knows only the popular news version of how these images are created.