If you're worried that our future robot overlords (or lovers) will not only enslave humanity but also say mean things, let this interesting tidbit put your mind at ease.
While developing Smart Reply — a new Gmail feature that employs a deep neural network (a teachable robot brain) to create auto-replies — Google researchers discovered that while learning the English language, the neophyte robot brains tended to gravitate toward expressions of love. According to Popular Mechanics, early Smart Reply prototypes would respond with "I love you" to just about anything.
And the honeymoon phase isn't over: Now, when the finished Smart Reply brain is left to "ramble" on its own, it still prefers saying things like "I love you, you're beautiful. I love you, you're beautiful" as a go-to response — like a boyfriend who really just wants a fight to end.
With that information, it's becoming clearer how in the future we'll go from a civilization that sort of fears robots to one that gladly accepts robot lovers.