Lawyer faces sanction for using ChatGPT in a client’s case

Attorney Steven Schwartz of the law firm Levidow, Levidow & Oberman used ChatGPT to prepare documents for a client\’s case. The lawyer “did not know” about the chatbot\’s ability to generate false content, writes The New York Times.

Levidow, Levidow & Oberman filed a lawsuit against the Latin American airline Avianca on behalf of Roberto Mata. He claimed that he sustained a knee injury after being struck by a metal service cart during a flight to John F. Kennedy International Airport.

When Avianca\’s lawyers asked Judge Kevin Castel to dismiss the suit, Mata\’s lawyers submitted a 10-page report explaining why it should be heard.

The document cited court decisions in several cases such as \”Vargese v. China Southern Airlines\” and \”Miller v. United Airlines\”. However, the judge and Avianca\’s lawyers were unable to locate the precedents cited in the report in the databases.

It emerged that Schwartz used ChatGPT to “augment” his case research.

 “I did not know that content [generated by an algorithm] could be false,” he wrote in an affidavit.

Schwartz also provided screenshots of the chat with the AI, where he asked the bot about the reality of the listed cases. ChatGPT replied that these precedents could be found in “reputable legal databases”, including Westlaw and LexisNexis.

According to Schwartz, he “deeply regrets” using ChatGPT. The attorney pledged he would never do this again in the future without full verification of AI content.

The judge called the situation “unprecedented” in which “the interpretation of the legal position consists of counterfeit decisions, citations and quotes”. He set a hearing for June 8, 2023 to consider “possible sanctions”.

Earlier in January, the AI startup DoNotPay offered lawyers $1 million for using an AI attorney in the U.S. Supreme Court.

That same month, the company\’s chief said professional associations in some states threatened him and his company with lawsuits. As a result, DoNotPay halted the use of the AI attorney in the upcoming trial.

In March, the startup was accused of illegal legal practice.