Video unix group

The nvidia tesla GNU/Linux documentation instructs admins on select distros to add users to the video unix group. I’ve seen this group mentioned for other Linux distros as well, even on AMD cards, but I cannot find detailed documentation on what it does and the implications of adding a user to it. A regular user being in this group doesn’t seem to be required in order to use CUDA or OpenGL in general.

Is there documentation on this group?