Citation and Attribution - Generative Artificial Intelligence - LibGuides at Brown University

Based on the Brown University LibGuide, the citation format that treats the AI tool as the author, includes the generation date, and allows for optional prompt descriptions in footnotes is Chicago Style.

• Chicago Style: Advises treating the AI tool as the author and the company as the publisher. It recommends including a note or parenthetical citation rather than a bibliography entry. If the prompt is not included in the main text, it can be added to a footnote or endnote.
• APA Style: Treats the company (e.g., OpenAI) as the author, not the tool itself. Prompts are described in the text (introduction or methods), not citations.
• MLA Style: Views AI as having no author. The citation starts with the prompt description in quotation marks as the title. Prompts are included in the Works Cited entry, not footnotes.

Brown University’s library guide on citation and attribution for generative AI provides a practical convention for documenting AI-generated material in academic work. Its central recommendation is to treat the AI tool itself as the author of the output, rather than assigning authorship to the human user, while recording the date the material was generated and, where useful, the prompt or instructions used to produce it. This approach is presented primarily through footnote-style citation, making it especially relevant for disciplines that rely on endnote or footnote conventions.

The guide’s key contribution is that it converts a still-ambiguous attribution problem into a concrete, reproducible documentation practice. By naming the AI system, specifying the generation date, and optionally including the prompt, the citation creates a minimal provenance record that helps readers assess the origin, context, and potential limitations of the material. For technically literate users, this matters because generative AI outputs can vary by model, prompt, interface, and time; the guide effectively treats the prompt and generation date as methodological metadata rather than ancillary details.

The material matters because it supports transparency, intellectual honesty, and auditability in an environment where the boundary between human authorship and machine-assisted production is increasingly blurred. Rather than requiring users to make ad hoc judgments about whether AI-generated text, summaries, or other artifacts require attribution, the guide offers a defensible baseline that can be adapted to different citation practices. It also helps institutions manage academic integrity concerns by encouraging users to disclose AI involvement clearly, while still distinguishing AI-generated content from human-authored scholarship.