Artificial Intelligence (AI) is often used to refer to a broad group of computer programs that are thought to be intelligent whether because they can generate predictions, recommendations, or decisions, or because they can approximate human thinking. “Generative AI” is commonly used to refer to a subset of AI programs that are designed to create, or generate, new content. That content could take a variety of different forms including new text (e.g., writing a new story or article), new images (e.g., creating a new painting or photograph), or new video (e.g., creating virtual video footage). Generative AI, in turn, can be further subcategorized along a variety of different lines. For example, some scholars and policymakers break the category into the following subgroups:[1]

1. General purpose AI (GPAI). GPAI refers to Generative AI designed to complete various functions. Functionally they can adapt to different requests from users.

2. Single purpose AI (SPAI). SPAI refers to Generative AI designed to complete a single specific task (e.g., draw a picture).

3. Large language Models (LLMs). LLMs are designed to recognize, predict, translate, summarize, and generate language.

[1] See Congressional Research Service, Generative Artificial Intelligence and Data Privacy: A Primer (May 23, 2023) at 2.