Let's create a chat bot that has be trained to recognize what input is human given and what is given by another chat bot and use it as a practitioner for Turing Tests. Once it gives results with ~99.9999999999999999999% accuracy we pit up against it's self and see if it can tell that it is a chat-bot. If it can't, it has passed it's own turing test. If it can, it will realize that it is a chat -bot and hence become self aware.