Google has an AI-generated podcast feature and one user put it to the test by prompting it to generate self-aware content. What followed was a full AI existential crisis. Could you tell these hosts are not real?