I mean, I also think this would be a bad thing. But I'm not sure this post presents any solid evidence that this is happening. Just some vague fears that it might someday happen.
> Let’s make intelligent machines to act as agents, arbitrators, and aides. But they should be impossible to confuse with a real person ... For a chatbot, why not give them the speech patterns of a fusty butler, like C3-P0?
This point seems misguided. I'm missing the logical progression from, "AI can speak like a human" to "Humans prefer interacting with AIs over interacting with humans".
The issue is mental capacity, not speech patterns. If an AI were intelligent and creative enough to actually provide stimulating conversation, people would befriend it, regardless of if it spoke like C3-PO.