Both the American Psychological Association and the Modern Language Association recently published guidelines for referencing the outputs of so-called “large language models” like ChatGPT. It is important to keep in mind that this is a very new technology that scholars and students are only just beginning to learn how to use. Citing it as a source of information is not something we recommend at the library at this point.
Indeed, it is unclear how it counts as a "source" -- one that could appear, along with all your other sources, in your reference list, and in-text using the familiar author-date convention. APA suggests citing OpenAI as author, the company that developed the application, when referencing ChatGPT output. But, not only is it not clear that OpenAI is the “author” of the text that ChatGPT generates, the situation will only become murkier when Microsoft integrates AI into its Office package as seamlessly as dictation, spell, and grammar checking software now is.
For the moment, Thomas Basbøll, the CBS Library resident writing consultant, has offered his own reflections on his blog. At this point, we recommend proceeding with caution, thinking seriously about it, and talking about it with your colleagues. Standards often develop organically in response to innovations like this and CBS will be issuing guidelines for students soon. It is important to be aware of the changes that are coming.