Let's just do a thought experiment here, OK?
Let's say your mom clicks on a malicious link. Let's say it's a drive-by. You know what that is. The computer can know, too.
Let's say that the computer mentally installs it and the computer predicts the consequences, kinda like looking ahead 1,000 moves ahead in a chess game.
Then, let's pretend that the computer knows that the intention of the operation is to load a keylogger, or it wants to make her computer a node in a botnet, or it wants to communicate with command and controllers "out there," ... or any other activity which YOU know is harmful.
If YOU know it's harmful, the computer can know as well.
That's the future. I'm very surprised it hasn't arrived way before now.
However, you didn't answer the first question: What is it you are going to do to your mom's computer that you can't get a computer to do for itself?