LexisNexis CEO says the AI law era is already here

October 27, 2025·1 hr 5 min·5 clips

LexisNexis CEO Sean Fitzpatrick didn't expect to hire an army of attorneys to review AI outputs, calling it their secret sauce.

Sean Fitzpatrick is the CEO of LexisNexis, one of the two dominant legal research database companies in the United States, and he speaks with Decoder host Nilay Patel about the collision between large language models and the legal profession. The conversation opens with the AI hallucination crisis in law: multiple cases have been documented in which attorneys submitted briefs citing court cases that do not exist, generated by AI tools that produce plausible-sounding but fabricated legal citations. At least two court rulings have been retracted after opposing counsel identified AI-generated hallucinated cases in submitted filings. Judges have sanctioned lawyers for this, and bar associations have begun issuing guidance. Fitzpatrick describes Protege, LexisNexis's AI legal research product, as built to avoid this failure mode by grounding every output in LexisNexis's verified database of actual court decisions, statutes, and legal commentary. He explains the architectural difference between a general-purpose LLM that generates text from statistical patterns and a retrieval-augmented system that cites specific verified documents. The product presents cited sources directly alongside AI-generated summaries, allowing attorneys to check the underlying text before relying on any claim. Fitzpatrick is candid that the hallucination problem is not fully solved even in his own product and describes the ongoing technical work to reduce it further. He also addresses the competitive dynamics with Thomson Reuters, which owns Westlaw — the other major legal database — and the race to integrate AI across the industry. He is blunt about the professional stakes: he believes it is a matter of time before an attorney loses their license due to AI-related negligence. The conversation also covers what law firms are actually willing to pay for AI tools and where the pricing models are settling.

As heard by us

A grounded look at LexisNexis pushing legal AI from research into court-facing drafting.

LexisNexis is trying to turn its legal research backbone into an AI drafting layer, and Nilay Patel presses Sean Fitzpatrick on the part that matters: what happens when Protege moves beyond finding cases and starts helping lawyers write the documents they file in court?

Read the full review in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Decoder with Nilay Patel