<\/p>\n
The media frenzy surrounding ChatGPT and other large language model artificial intelligence systems spans a range of themes, from the prosaic \u2013\u00a0large language models could replace conventional web search<\/a>\u00a0\u2013 to the concerning \u2013 AI will eliminate many jobs \u2013 and the overwrought \u2013 AI poses an extinction-level threat to humanity. All of these themes have a common denominator: large language models herald artificial intelligence that will supersede humanity.<\/p>\n But large language models, for all their complexity, are actually really dumb. And despite the name \u201cartificial intelligence,\u201d they\u2019re completely dependent on human knowledge and labor. They can\u2019t reliably generate new knowledge, of course, but there\u2019s more to it than that.<\/p>\n ChatGPT can\u2019t learn, improve or even stay up to date without humans giving it new content and telling it how to interpret that content, not to mention programming the model and building, maintaining and powering its hardware. To understand why, you first have to understand how ChatGPT and similar models work, and the role humans play in making them work.<\/p>\n Large language models like ChatGPT work, broadly, by\u00a0predicting what characters, words and sentences<\/a>\u00a0should follow one another in sequence based on training data sets. In the case of ChatGPT, the training data set contains immense quantities of public text scraped from the internet.ChatGPT works by statistics, not by understanding words.<\/p>\n \nHow ChatGPT works<\/h2>\n