While AI language generation models like GPT-2 are impressive, the statistical approach to passing or fooling a human in a conversation is not enough to fully solve the Turing test. Despite competitions and advancements in language understanding and generation, there are still several rules to be addressed in order to achieve a natural conversation.