In Mississippi, a lawyer included AI-generated fake citations and was told not to do that, but kept doing it.
The latest judge to receive her fake citations was not amused, and has issued sanctions against her and the two partners in the firm that she worked for.
If I’m understanding right, the partners have now dissolved the firm.
(It looks like Ms. Watson, the lawyer who used AI-generated fake citations in ten different cases, may be the daughter of one of the two partners.)
The judge reacted strongly to Ms. Watson’s behavior:
In light of repeated warnings from federal courts about the risk of hallucinated cases, as well as CLE trainings she attended, direct notice and knowledge of the same prior mistakes, her violation of the Firm’s AI policy, and the sheer number of filings, Ms. Watson’s misconduct is particularly egregious and prolific.
The partners are also being sanctioned for failing to notice the problems. For example:
a large portion of Billups’ argument relies on a case styled Jackson v. Gautreaux, 3 F. 4th 182, 190 (5th Cir. 2021). […] In fact, this case is cited eight times, even arguing that a jury should be instructed under its holding. […] In reality, Jackson is an excessive force and failure to train case and is wholly irrelevant to the case at bar. A seasoned attorney examining the brief should have read a case so heavily relied upon. Had he done so, he would have easily discovered the problems.
The judge noted that the usual penalty for this sort of thing has been fines, but quoted another case about why fines are insufficient:
“If fines and public embarrassment were effective deterrents, there would not be so many [AI misuse] cases to cite.”
(Given that there are so many such cases, I probably won’t post about all the ones I hear about, but this one did seem especially egregious.)