
The California Court of Appeal upheld a family‑court order that relied on two nonexistent cases, *Marriage of Twigg* and *Marriage of Teegarden*. The order was drafted by the appellant’s counsel, who failed to verify the citations, while the opposing attorney was sanctioned $5,000 for introducing the fabricated authority. The court emphasized the forfeiture rule, finding that the appellant’s lawyer forfeited any objection by submitting the flawed order without protest. The decision highlights the growing risk of AI‑generated “hallucinated” case law in legal practice.
The *Torres v. Munoz* opinion illustrates a new frontier of legal malpractice: AI‑generated hallucinations. In this case, counsel relied on a Reddit‑sourced article and an unnamed AI assistant to cite the fictitious *Twigg* and *Teegarden* decisions. When the appellate court discovered the citations were invented, it imposed a $5,000 sanction on the attorney who introduced them, yet it affirmed the underlying custody order because the appellant’s own lawyer had drafted and filed the document without flagging the error. This juxtaposition of sanction and affirmation highlights how courts balance procedural fairness with the need to deter careless citation practices.
The decision reinforces a long‑standing ethical obligation: lawyers must conduct a reasonable search to confirm the existence and relevance of any authority they cite. The forfeiture rule, applied here, bars a party from challenging a court’s reliance on a citation when that party had the opportunity to object but failed to do so. Practitioners are reminded that drafting a proposed order carries the same verification duty as filing a brief. Ignoring this duty not only risks sanctions but can also jeopardize a client’s substantive rights if the court’s reasoning is later called into question.
For law firms, the case serves as a cautionary tale about integrating generative AI into research workflows. While AI can accelerate document review, it can also fabricate citations that appear plausible. Best practices now include a mandatory double‑check of any AI‑suggested authority using official reporters or subscription databases, and documenting that verification in the record. As courts become more vigilant, firms that adopt rigorous verification protocols will protect both their reputations and their clients from costly procedural setbacks.
Comments
Want to join the conversation?