I think my concern is that if you talk to a bot in a certain manner, it normalizes talking to humans in that manner, if the bot is "human enough". I wonder if

I think my concern is that if you talk to a bot in a certain manner, it normalizes talking to humans in that manner, if the bot is "human enough". I wonder if the solution is to stop building bots to emulate humans: Natural language processing is important so humans can make ordinary requests of a computer without having to remember specific syntax. But perhaps we should stop trying to make the computer respond the way a human would, so that people don't view it like a human or personality.