Users submit an application to vote remotely through a government-controlled site and confirm it by receiving a SMS with code. It is obvious that government is able to submit such an application to its own server without bothering the real user.
The users vote remotely and confirm their identity by providing a code from SMS (again) and the government can just look at the code on the server. It isn't important what cryptography is used there. You don't need to guess random numbers when you have root access to the server.
The observers are supposed to see a tape where some hashes will be printed when someone votes. Maybe they will be allowed to see the registry of voters who chose to vote remotely, maybe they will not be allowed.
I cannot prove this, and this is my personal opinion, so you shouldn't believe me, but I think the government wants an opaque procedure to produce legitimate looking election results with desired outcome. With paper voting, there are too many points of failure: it is difficult to throw the ballots in the box when there is a vigilant observer. Remote voting doesn't have such flaws.
If there was a fraud, it would be difficult to explain to a non-programmer. They don't understand anything, they just see how the program says "hash doesn't match". But maybe it is just a program written by foreign agents (and stored on foreign "Github" servers) producing fake results. For comparison, with paper voting anyone can understand that there is a fraud if you show them the video.
Also, the government party is slowly losing its rating and starts losing regional elections. Remote voting might be a cure for this undesired situation.
And probably any other government (for example Estonian) that implements such an opaque system has similar aims.