It's just an advanced chatbot that reads fake news online and uses it as a source for the wrong answers it gives you. You can easily show it when it's wrong and it will learn in that one session, but it doesn't have the knowledge or discernment to understand that it's pulling up bad sources. It has no intelligence, it's just an aggregator that delivers to you in conversational style. Really good prompt-crafting helps you avoid these issues. Remember it's new tech, it's going to have lots of hiccups in the beginning, this is especially evident in controversial topics.