By the way, another possibility is to activate the root account and use it only in a separated TTY. So e.g., CTRL+ALT+F5 to switch from the GUI’s TTY to a normal terminal and than use only this one for logging in as root.
Formally, if you would use the same password for both (user+root), and if you log into both only directly (so not from within one account logging into the other), the user accounts themselves would never come in contact to your password, because if you only use it to log in, then it is the system that checks if the password matches, and if so, it logs you in as the respective user. Thus, the latter never comes in contact to with the actual password or its hash in /etc/shadow.
This means, you would have a separation, and could use the root account only when necessary for relevant actions.
Generally, this is not considered a good practice because it adds a strong social vulnerability: users tend to enter their password when they are asked for. We do this often intuitively. Especially when people are used to do admin stuff from within the account. Thus, if the GUI would be captured or leaked, the attacker would have also the password for root. This could thus bypass the separation by social means (in case you introduce this to prevent exploitation of captured / leaking GUI through malicious or vulnerable apps).
Anyway, separating root and the user account (at the best, as good practice, with separated passwords) is one way to avoid the user account to become the possibility to act as superuser. And to exclude any connection between GUIs and superuser.
With user_u (SELinux confinement) you could even forbid the user to even try su/sudo, but that does currently add some restrictions that break some functionalities and thus is not yet eligible for production use.
