Generative AI creates output rather than selecting among existing items. In legal work this means drafting, summarizing, answering questions, and restructuring information, as distinguished from classification tools that sort documents or search tools that return them.
The distinction matters because generative output is synthesized text that may not correspond exactly to any source unless the system is designed to constrain it.
Alternative Names:
GenAI, Generative Artificial Intelligence
Why it Matters?
The generative and classification distinction determines how output must be checked. A classifier that mislabels a document produces an error a reviewer can catch by reading it. A generative summary that subtly misstates a fact reads fluently and correctly, which is precisely why it is harder to catch. Verification workflows should be designed around that asymmetry.
Frequently Confused with
Related terms
Frequently asked questions
How is generative AI different from the AI already used in e-discovery?
Why do court orders single out generative AI?
