All photos are basically rotated, more precisely inverted. This is because the display system of the camera/mobile phone has to create an actual image on the sensor. If we want to use a lens/mirror to create a real image, it will always be inverted. In the same way, the lens in our eye creates an inverted image on the retina and the brain “turns” it around.

So an inverted image is created on the phone’s sensor and the phone automatically rotates it so that the image looks the way we see it on the display. So the phone rotates the image before displaying it but only vertically, but there is no fundamental problem to rotate the image in any other way. So when taking a selfie, the image can be both rotated and mirror inverted. And the word mirror is, in my opinion, the reason why selfie photos are inverted — so that when looking at the phone’s display, we perceive ourselves as if we were looking in a mirror. This is the view we are used to, which is why the camera also simulates it. It’s just one extra operation for the phone’s hardware but it’s more comfortable for the photographer.

If we wanted to avoid rotating the image when taking a photo, we could use two lenses instead of one — each lens would invert the image, so after two passes the image would be upright. But this would complicate the construction of the camera and the optical system would also be more bulky.