You're right about the physical / optical limitation. There is only one plane of focus, and you can't use multiple focus-system focus points to get around that.

The advantage is simply that you can choose what to focus on without changing the framing or otherwise recomposing. If the thing you want in focus is over in a corner, you can select the corresponding focus point, and there you go.

Additionally, modern advanced autofocus systems generally have modes where they track subject motion. This is super-useful for sports, wildlife — or just your kids at the playground. More focus points (and ideally a wide spread of them) make this possible.

Modes where the camera allows more than one focus point to be active do one of three things (depending on the specific camera and settings):

1. They grab one of the focus points as determined by the camera's processor. Probably the one that's most in focus already, or the most contrasty — but it could be more complicated than that.
2. They try to optimize depth of field so that multiple points are in as much focus as possible.
3. They use the multiple points as part of a tracking mode, as above.