On the reliability of Large Language Models to misinformed and demographically informed prompts | Publicación