I have set up Google Chrome as a Remote Application on a Windows Server 2008 R2 machine. I can then load the application on another (Windows 10) machine outside the LAN and browse as if I was doing it from the host Server 2008 machine. It kinda acts as a proxy machine. The whole idea is to protect the client machine (Win 10) from anything nasty that could come through the browser and keep it at the host machine. The host machine is just a RDP enabled machine and I am less concerned about it's security than that for the client machine. Is this a safe practice, or could malware on the host machine somehow compromise the client machine when connecting to the remote application?
Security is rarely ever absolute. I mean it in the sense of not being able to say "this is absolutely safe" (negative assertions are somewhat easier).
That said, YES, using the remote app increases the effort for an attacker to compromise your client machine and in that sense, it is more secure than not using a hop.
Know though, that once the attacker compromises the remote machine it is only a matter of time before they get to your client system through the RDP server. Granted, it is not going to be straight forward, but having access is the first step. Even if they don't compromise your system, your data (while you are browsing) is up for grabs.
I would say its safe, at least no infection could be done through the RDP itself, technically no malware can spread through this protocol. BUT you should check what you did allow once the RDP connexion is established like shared clipboard or redirected disks which you can disable or limit either through group policy or simply under Local Resources in the Properties of the RDP Client.