Midpage test finds AI web search research cited cases it never opened
Midpage said Nov. 6 that it ran the same AI model, Claude Opus 5.5, on 100 legal research questions twice: once with Midpage's case law tools and once with ordinary web search. In Midpage's test, the web search version had opened only 33% of the cases its answers relied on, compared with 82% using Midpage, and an AI judge preferred the Midpage answer on 89 of 100 questions. This is the company's own test, graded by an AI model, not an independent study. Whatever tool you use, open and read each case an AI answer cites before you rely on it.
Source: midpage.ai (opens in a new tab)