Generative AI models predict the next token (or pixel) to produce new text, images, or code. They can be fluent and still be wrong. Treat the output as a draft, not as an official circular, a medical opinion, or a live gold rate.
If you use these tools for exam prep, office mail, or code, you still need to check facts against a primary source. Janooji's job is the source-backed answer, not an unverified chatbot paragraph.
