That's not how that term works. Input latency/lag refers to the time differential between triggering some kind of input and the action of that input showing up on your screen.
I've got my homelab participating in Folding@Home, it'd be interesting to run some commercial workloads in a similar manner to help offset energy costs.
There are two levels of AWS Shield, "Standard" and "Advanced". You are correct that all instances receive "Standard" protection from Shield by default.