The use of ChatGPT and artificial intelligence in the legal sector has raised questions about its reliability and the need for human oversight. One of the most recent cases highlights the story of a lawyer who used ChatGPT to gather information for a legal case, but faced unforeseen consequences.
The challenge of using ChatGPT in the legal sector
The case in question involves attorney Steven Schwartz of the law firm Levidow and Oberman. Schwartz decided to use ChatGPT to conduct in-depth research on cases similar to that of his client, Roberto Mata. The customer had sued the Colombian airline Avianca following a knee injury caused by a trolley during a scheduled flight. However, the use of ChatGPT has resulted in erroneous and fabricated information, which has been presented as evidence in the legal case.
Using ChatGPT in a legal case
Attorney Steven Schwartz decided to use ChatGPT to convince the federal judge not to dismiss the case of his client, Roberto Mata. However, using ChatGPT has led to misleading results and unforeseen consequences. Schwartz entrusted ChatGPT with the search for cases similar to that of his client. Using the chatbot, he obtained a document containing a list of alleged cases, including “Varghese v. China Southern Airlines”, “Martinez v. Delta Airlines” and “Miller v. United Airlines.”
Although the lawyer asked ChatGPT if the information was true, the chatbot wrongly confirmed their authenticity, stating that they were present in reliable legal databases such as Westlaw e LexisNexis. Subsequently, the defense attorney carried out a thorough verification of the cases cited by ChatGPT. It turned out that none of these cases actually existed. For example, the case “Varghese v. China Southern Airlines Co” turned out to be non-existent, despite referencing the actual case of “Zicherman v. Korean Air Lines Co.”.
The consequences for the lawyer
The use of inaccurate information in the legal process seriously undermined Attorney Schwartz’s credibility. His affidavit, in which he said he asked ChatGPT if he was lying, was not enough to mitigate the legal consequences of false information presented in court. The lawyer could face fines and Mata’s case could be compromised due to the incorrect use of artificial intelligence. In fact, the responsibility for what happened lies with the lawyer, who did not correctly use the information provided by ChatGPT.
The submission of incorrect information had consequences for attorney Schwartz. He was arraigned by a federal judge for presenting bogus evidence. Despite his pleas of ignorance about the possibility that the information was false, he was found guilty.
The importance of human supervision in using ChatGPT
The story of the lawyer Schwartz highlights the importance of careful and responsible human supervision in the use of ChatGPT. This is especially true in the legal sector. Generative AI can certainly be a useful tool for information research and processing. But it is crucial that lawyers are aware of the limitations of these technologies. Lawyers are expected to use information ethically and responsibly. Rigorous human oversight is essential to avoid errors, false information and unintended legal consequences.
The story of the lawyer Schwartz represents an important lesson on the use of ChatGPT in the legal sector. It is crucial that attorneys understand the limitations of generative AI and are responsible in using those tools. But this generally applies to all sectors that make use of the support of artificial intelligence. We must not forget that artificial intelligence is only a precious ally, but it cannot replace human intelligence. This story proves that human guidance and control are essential to avoid any unfortunate consequences.
Master’s student in Energy Engineering at the University of Bologna, always passionate about scientific dissemination, especially everything related to the environment and sustainability. Writing has given me the opportunity to meet new worlds and discover new interests, which is why despite having started as an author for #EnergyCuE since November 2020, today I collaborate for the entire network.