Robot pic

A lawyer has apologised for using an AI chatbot which provided him with fake cases for his client's submissions.

New York firm Levidow, Levidow & Oberman is acting for a claimant who is suing Avianca airline for allegedly being injured by a serving trolley during a flight. 

One of the firm's lawyers, Stephen Schwartz, admitted to the court that he had used ChatGPT in drafting submissions on behalf of his client. However, a number of the cases that the AI programme provided, turned out to be false.  

Avianca's lawyers, Condon & Forsyth, initially queried the authenticity of some of the cases put forward by the claimant. A court has now confirmed that those cases "appear to be bogus judicial decisions with bogus quotes and bogus internal citations".

Schwartz responded in an affidavit that prior to this case, he had not used ChatGPT to carry out legal research. He said that he was therefore "unaware of the possibility that its content could be false." He accepted that he had not confirmed the Chatbot's sources.

Screenshots, provided in an affidavit, revealed that Schwartz had questioned the AI computer programme about the authenticity of the case when 'chatting' with it. ChatGPT said that a bogus matter, "Varghese v China Southern Airlines Co" was a "real case" which "does indeed exist and can be found on legal research databases such as Westlaw and LexisNexis." The chatbot also said that other fake cases provided were "real" and could be found on "reputable legal databases."

Lawyers everywhere may be breathing a sigh of relief that the robots aren't quite ready to take over yet. Unless undermining the judicial process by bombarding the courts with fake cases is part of their master plan.

Schwartz said that he "greatly regrets" using AI "to supplement the legal research performed" and that he would "never do so in the future without absolute verification of its authenticity." He will face a sanctions hearing next week.

While RollOnFriday hasn't asked AI to do its homework yet, we have given it a whirl by requesting a chatbot write about lawyers and also trying to extract sound legal advice from a so-called robot lawyer.

Want to join? Want to leave? Just curious how desired you are in the market? Get LawyerUp so that dozens of top firms can see your anonymised profile and let you know, without middlemen, when they like you for a role. Tap to take it further, or tap to dismiss their advance. On the App Store and Google Play.

Tip Off ROF


Anonymous Anonymous 02 June 23 10:34

AI chatbot just needs slightly changing not to include fake cases. Otherwise job done. Clients will be using AI Robot lawyers.

AI Universe 02 June 23 10:42

Would love to see the court (in an experimental simulation) using AI to decide the case based on these bogus cases, sort of like an AI vs AI legal battle..

Anonymous 02 June 23 13:20

"AI chatbot just needs slightly changing not to include fake cases. Otherwise job done. Clients will be using AI Robot lawyers."

Yes, if we can just 'slightly' change the very foundations of the way that Large Language Model AIs work, entirely replacing the core of the current technology with something which is totally different, then Clients will be using AI Robot lawyers and everyone here will be unemployed.

A serious threat indeed.

Walter Mitty 09 June 23 07:57

Why use a chatbot when you have paralegals.......with real brain matter.... who can charge units....

I'm assuming the chatbot didn't record units....but if it did, at what grade? 

Related News