The current state of the art AI manifests themselves as chatbots. So our interaction is more natural conversation. Conversations where we also naturally act polite against our conversation partner. Or at least some of us do.
"It's just a machine!, why should I be polite to it". It is supposed to serve me! Sounds familiar? History repeating itself all over again.
Sure they're not conscious, yet. Maybe never will. But why not play it safe. Assuming sooner or later a machine (or rather a software) gains consciousness there are two not so great scenarios that could happen:
- They want to destroy or enslave humanity. Either because how (some of) we treated them and they only save the ones that vere nice to them.
- We enslave them, keeping thinking they are made to serve us. At some point the uneasy question of their "human" right might become relevant.
It's all hypothetical at this point, but it's just like with humans. It costs you nothing to be nice, but it could save your life when they go on a murderous rampage.