In a case of ‘putting lipstick on a pig’: Lawyers using ChatGPT for legal work

ChatGPT – A generative language processing tool, not a search engine

Everyone is talking about the ChatGPT which is an artificial intelligence chatbot developed by Open AI which launched on 30 November 2022. ChatGPT is not a search engine, but rather a generative language processing tool. It is trained to generate words from given input, however it is not capable of comprehending their meanings. In other words, it is not a “BS” generator.

We have with much amusement been experimenting with ChatGPT – the fake legislation and hilarious but phony case names – it is actually awe inspiring how authoritative it sounds.

Incredulously students and politicians have openly admitted to using AI in their work, without too many concerns are that AI is producing false information to users. And only in the US fashion – US lawyers were fined for citing fake cases using ChatGPT in submissions for a personal injury case.

Key Facts

Let’s have a closer look – Avianca Airlines, a Columbian-based airline, sought to dismiss a federal court case where an injured passenger claimed he was struck by a metal serving cart on a 2019 flight.

When preparing a response, the lawyers relied on as precedent and cited cases such as Varghese v China Southern Airlines and Shaboon v Egypt Air (told you they sounded convincing). The court found that these cases did not actually exist. Although ChatGPT assured the lawyers that these cases were real, they were in fact fake.

Now we have heard of lawyers failing to check a footnote – but the lawyer who used ChatGPT admitted that he used the AI chatbot, but he argued that he should not be sanctioned because he had no intent to deceive the court and did not act in bad faith.

Judge P. Kevin Castel found that the lawyers had acted in bad faith and made “acts of conscious avoidance and false and misleading statements to the court.”[1]

The judge said there was nothing “inherently improper” about using a reliable artificial intelligence tool in assisting with legal work, however lawyers need to ensure their fillings are accurate.[2]

Key Takeaways

In general, lawyers need to move on from questioning the polish-ability of turds sorry words. The key takeaway from this case is to use AI chatbots with caution.

[1] Mata v. Avianca, Inc., No. 1:2022-cv-01461 (S.D.N.Y., 2023).

[2] Ibid.