Summary
In *Hill v. Workday*, Magistrate Judge Peter H. Kang sanctioned a supervising partner, not just the associate who filed an AI-tainted brief, after finding that the firm's internal training and review practices did not excuse the partner's failure to verify a non-existent case citation and a misleading case summary. The order is a strong direct legal record because it moves the AI-misuse story from individual drafting error to managerial responsibility, remediation, and training.
Why It Matters
This matters directly for lawyers operationalizing AI because the order spells out where responsibility sits when AI-assisted work goes wrong:
- a partner cannot satisfy Rule 11-style duties by assuming that firm training or junior-lawyer review is enough
- supervisory responsibility includes personally reading the filing and understanding the cited authorities
- AI governance inside a firm has to be procedural, not just cultural or educational
- courts may impose tailored remedies such as CLE, internal circulation, and monetary sanctions rather than stopping at public embarrassment
What the Source Says
The April 28, 2026 order says attorney Katherine Cervantes had already been sanctioned in September 2025 for submitting a brief that cited a case that did not exist and misstated another authority after using generative AI. The follow-up order explains that supervising lawyer Lenden Webb still signed and filed the brief even though he had not checked the authorities personally. Judge Kang wrote that a lawyer's obligation to certify a filing cannot be met "without reading the filing, understanding the arguments, and verifying the authorities that support it." The court rejected Webb's reliance on firmwide AI training and weekly cross-training because those measures did not create a specific personal verification step before filing. As sanctions, the order required Webb to pay $1,001, complete four hours of live CLE split between AI ethics and supervisory duties, circulate the sanctions orders internally, and file proof of compliance.