AI does not think, it uses its training data to generate predictive text, suggestions, or solutions.
When you text, the predictive text is based on what the app knows about you and the words that make sense next.
Generative AI works in a similar way – it uses MASSIVE amounts of training data and the rules that have been created for it, to generate sentences and paragraphs that are meant to answer your questions.
Most AI are trained on data-sets which consist of text or images from the open web, places like Reddit and Wikipedia.
Many new tools and technologies are using base models like ChatGPT for their services.
Large Language Models are not designed to know where they got their information and often generate information that is incorrect.
Often generative AI will cite a "hallucination." This may have a real author and a real journal, but the article it is citing doesn't actually exist. Hallucinations can also refer to Generative AI saying something that is not accurate.
Sometimes Generative AI will link to a citation that is real, so you must always check to verify that the citation supports the generated statement