A legal brief can sound convincing and still cite cases that never happened. Accuracy means that an output is correct for the task at hand. For a large language model (LLM) drafting a brief, fluent wording does not establish that its legal claims are accurate. section 5.8
In Mata v. Avianca, an attorney used ChatGPT to prepare a brief without checking its claims against reliable sources. The incident illustrates a practical limit: LLMs can produce plausible prose, but they cannot generally ensure factual accuracy. Verification matters especially in work that requires specialized knowledge, including law, medicine, and finance. section 5.8
