Generative AI

Generative AI is AI that produces new content, such as text, images, audio, video, or code, rather than just analyzing or classifying material that already exists. The text kind is built on large language models, and tools like ChatGPT and image generators are its most familiar examples.

The distinction matters here because much of what gets lumped under "AI" is not generative. Lower-risk language technologies like OCR, text-to-speech, and spellcheckers convert or analyze existing material rather than inventing new content, and they often get swept up in criticism aimed at generative systems (see Natural language processing). Generative AI carries sharper risks precisely because it makes things up: being probabilistic, it can produce fluent misinformation, misrepresent a language or person, or put words in someone's mouth that they never said (see Ojibwe Chat and Chatbots with Human Likeness).

Created · Updated
Supported By the National Science Foundation Award 2542375.