"The problem is to somehow convince most people that you are really a chatbot and not a human."

Er. Is that seriously a problem? I talked to one of the chatbots for about 15 seconds and virtually every response I got was poor

But suppose you get a few responses that seem surprisingly good. You might think that there's a human watching the chat and responding every once in a while.

If you don't have an interest in chatbots, consider another example: targeted advertising in gmail. How do you convince people that there is no human involvement in this targeting at all?