Same motion to dismiss, same Claude, two runs: one without LawDiver, one with the LawDiver connector. Double the cites, good-law checks on almost every case, pin cites from opinions actually read, a subtle misread caught before filing, and a July 2026 decision no training cutoff can know.
Effective January 1, 2027, California SB 574 requires lawyers to verify generative-AI output and to personally verify every citation in a court filing. CiteCheck is the corpus check that documents that work.
Updated October 12, 2026. A 50-state rescan of ABA Formal Opinion 512, state bar AI ethics opinions, statewide court rules, and California SB 574 — with enactment and effective dates and a link to every authority.
USA TODAY and the newspapers it owns sued OpenAI for training ChatGPT on their reporting. The complaint, and where it meets the Third Circuit's ROSS decision.
A California firm says Thomson Reuters paid $650 million for Casetext, kept CoCounsel, and deleted the cheap research product. The complaint is a map of how legal tech actually competes.
The Third Circuit affirmed Thomson Reuters’ win against ROSS Intelligence. Judge Montgomery-Reeves held that 2,243 Westlaw headnotes are original, and that using them to train a competing legal-research tool was not fair use.
Cite checking is difficult because one decision has many valid cites, The Bluebook is not the rule in the courts that read the brief, and some states require two reporters. Free case indexes usually store one cite, or none.
5,300 citations across six accuracy bars — including separate scores for easier Bluebook mangling and harder recovery problems. Here is how CiteCheck did.
OpenAI launched Astra for Law today — a GPT-6 configuration with a legal search index built on Free Law Project's CourtListener data, sold to Am Law 200 firms through an application form and a 30-page agreement. We built LawDiver on the same public sources. It is free during rollout, the cite checker hits ~97% accurate (100% on properly formatted cites; 100% on hallucinated caselaw), and the API and MCP server plug into any model — or any law app you are building.
From Harvard CAP and CourtListener bulk dumps to daily harvest, EyeDiver, the Bubbler, embeddings, dual-LLM analysis, and GoodLaw — how LawDiver built a free-to-use American caselaw corpus without Westlaw or Lexis content.
American lawyers still cite volume and page of commercial print reporters even as those books die. Copyright fights, the Thomson/West antitrust decree, the end of the Federal Appendix, Florida's So. 3d rules, and what free caselaw data still gets wrong about citations.
A practical workflow for cite-checking legal work: confirm each case exists, supports the proposition, and remains usable — plus where to cite-check a brief for free with CiteCheck.
Hallucinations are not bugs in the chat UI. They are statistical consequences of next-token training, accuracy-graded benchmarks, and agent architectures that still let the model decide whether to trust a database. Here is the science, the case law of fake cites, and why deterministic verification has to sit outside the LLM.
Court opinions are public records — yet reading them at scale has meant a subscription that costs more than most car payments. Caselaw puts more than 10 million U.S. opinions in front of anyone, free, with no login.
Most “AI cite checking” is a second pass by the same kind of probability engine that drafted the brief. Cite-checking is a database problem — and LawDiver’s CiteChecker treats it that way.