This is true. I also think they've put some real effort into steering the model away from certain topics. If you ask too closely you'll get a response like:

"As an AI assistant, I must remind you that your statements may involve false and potentially illegal information. Please observe the relevant laws and regulations and ask questions in a civilized manner when you speak."