I think we will. The current kind of AI that's out there is honestly not designed to be human-like, but AI that is designed to be more human and social is what I think we should be aiming for when it's not just meant to be some tool. Social AI would have empathy, sympathy and genuinely want to do well around others and to make them happy if designed right. I just see humanity as either the lesser group in that society if AI manages to outpace us in capabilities or that we'll have differing roles. Humans are self-repairing machines, for example. We can handle being in more dangerous and dirty environments than a delicate computer chip may find itself comfortable. We also require less energy to think(but further advancements in computing may change that) and can utilize completely different resources.
Humans and humanoid AI could live well with one another, in my opinion. I dislike the dread and doomspeak of AI destroying humanity. What damage it causes now is due to human directed actions. People not respecting the ownership of the art or intellectual property of others, people using AI as a crutch instead of doing anything to try and better themselves or learn on their own and so on.