Are we seeing the case where AI is now suffering from multiple personality disorder? As much as fascinating this is, I think the fact that an LLM cannot _really_ think for itself opens it up to abuse from humans.