Automated summarization produces shorter versions of source material. Extractive approaches select existing sentences, while abstractive approaches generate new text, which reads better but introduces the possibility of statements not present in the source.
Quality is judged by faithfulness to the source as much as by brevity.
Alternative Names:
AI Summarization, Document Summarization
Why it Matters?
Summarization is the most widely used AI capability in legal work and the one where errors are hardest to notice, because a fluent summary carries no signal that something was dropped or shifted. The controls that matter are source citation at the passage level and a habit of spot-checking against the original on anything consequential. Summaries of documents that will be relied on for advice or filings warrant reading the underlying source.
Frequently Confused with
Related terms
Frequently asked questions
What is the difference between extractive and abstractive summarization?
How should summaries be checked?
