The force reboot part sounds particularly lame to me.
For server SKUs, each process has right to delay shutdown for 5 minute, they may get away from it with clean shutdown. Since Vista each process on Consumer SKUs has only 2 seconds to do cleanup. That means applications with lots to write when quit will do incomplete write on shutdown, and with transactionaly write on NTFS, it means your previous work is gone.
Realistically, I think it's the opposite. Servers have much more exposure to hackers than ordinary users. For example, it took hackers about 4 hours after Heartbleed was announced to have scanned the entire IPv4 range for servers that were vulnerable and begin exploiting them.
Servers also have no excuse for not building in redundancy. If your server can't handle an unexpected reboot, it can't handle a server crash either. All servers realistically should be at least pairs of machines as a clustered availability set so that one can handle the requests when the other goes down. In such a scenario, Windows Updates hard-rebooting the machine isn't even a problem so long as it doesn't hard-reboot all of the machines in the availability set all at the same time.