AI Hallucinations in US Courts Increase Nearly Sevenfold — Two Attorneys Disqualified for Two Years in June
A new study found that US court matters tied to AI hallucinations rose from 230 a year ago to 1,667 by mid-2026 — a nearly sevenfold increase. Leading AI legal tools including Lexis+ AI and Thomson Reuters systems have been benchmarked with incorrect information rates exceeding 17%, with some assessments pushing past 34%. In June 2026, a US District Judge disqualified two attorneys for two years a
BY FRONTIER DESK · JULY 7, 2026 · 1 MIN READ
A new study found that US court matters tied to AI hallucinations rose from 230 a year ago to 1,667 by mid-2026 — a nearly sevenfold increase. Leading AI legal tools including Lexis+ AI and Thomson Reuters systems have been benchmarked with incorrect information rates exceeding 17%, with some assessments pushing past 34%. In June 2026, a US District Judge disqualified two attorneys for two years after they submitted AI-generated legal research containing hallucinations. Sullivan & Cromwell admitted in April 2026 to a court that a bankruptcy filing included fake citations created by AI. Damien Charlotin's global database of AI-related legal misconduct now tracks more than 1,600 incidents worldwide. For mid-sized firm managing partners, the risk profile is now clearly defined: the "grace period" for learning through AI errors has ended, judges are imposing two-year disqualifications, and a mandatory human verification step for every AI-generated citation is not optional — it is the baseline liability management requirement. Structured training for lawyers, paralegals, and associates should be treated as a governance control, not a technology orientation.