Advocates who thoughtlessly applied AI were fined in the USA

The number of cases where advocates use artificial intelligence without proper verification is growing. This trend is alarming to judges, as AI hallucinations can find their way into court decisions.

Federal Judge Michael Wilner fined two law firms, K&L Gates and Ellis George LLP, $31,000 for submitting a document to the court with fictitious legal references and citations created using artificial intelligence tools.

As part of providing legal assistance to a client in a civil case, an advocate from one of these firms prepared a draft submission. Another law firm included this information in the final version of the document and submitted it to the court.

While reviewing the submission, the judge became interested in the references and citations in the text.

After checking them, he found that out of 27 references in a ten-page document, 9 were incorrect and 2 were completely fabricated. Accordingly, the citations were also false.

After discovering the problem, the judge asked for an explanation. The advocates confirmed under oath that they had used AI (specifically Google Gemini and Westlaw Precision with CoCounsel).

The judge then stated that no lawyer or employee of the firms had apparently checked or reviewed the research before submitting the document. He also noted that this approach had almost led to a more dangerous outcome: the inclusion of false information in a court decision.

In imposing financial sanctions on the firms, the judge stated that no competent advocate should outsource research and writing to AI.