Stanford researchers benchmarked these platforms and found significant hallucination rates: Lexis+ AI and Ask Practical Law AI produced incorrect information in more than 17% of test queries, while Westlaw's tool exceeded 34%. The errors range from fabricated case citations to unsupported legal claims that appear credible on their face. The problem has accelerated sharply. Researcher Damien Charlotin documented filings containing hallucinated AI content rising from approximately two cases per week in early 2025 to two or three per day by spring 2025.
Attorneys should treat AI-generated legal research as unverified draft material requiring independent verification before any filing. Arizona courts and bar authorities now explicitly mandate this practice. The trend creates dual exposure: direct sanctions risk for lawyers who submit unvetted AI output, and potential product liability claims against vendors whose tools generate false citations. As courts treat these errors as professional responsibility failures rather than software limitations, the distinction between using AI as a research aid and relying on it as a substitute for legal judgment has become a critical line in malpractice and ethics disputes.