The modern-day lawyer utilizes technology in every aspect of their practice. Most lawyers communicate primarily—if not exclusively—with clients and opposing counsel through email. Microsoft Teams meetings and Zoom court appearances are now everyday occurrences. Some lawyers even hire virtual assistants who live overseas to assist in running the law practice. The latest technological advance to arrive on the scene are artificial intelligence (“AI”) chat bots which can respond to prompts with specific information and answers. Technology is here to stay, and the modern-day lawyer must understand the risks and benefits of relying on technology in their law practice.
A Quick Fix: Chat GPT
The story of Steven A. Schwartz, a New York lawyer who relied on Chat GPT to prepare a court filing, is an example of how using technology in the practice of law could go horribly wrong. Mr. Schwartz represents plaintiff Roberto Mata in a personal injury action against the defendant, an airline named Avianca, pending in federal district court in the Southern District of New York. Avianca moved to dismiss the lawsuit. In preparing an opposition to the motion to dismiss, Mr. Schwartz utilized the artificial intelligence website Chat GPT to provide case law to support his argument. Chat GPT provided Mr. Schwartz with the legal citations and assured Mr. Schwartz of their reliability. He then incorporated these cases into his brief. Mr. Schwartz did not double check the AI chatbot’s legal research instead submitting the brief based on the research done by Chat GPT.
Counsel for the defendant airlines soon approached the court and informed the judge that they were unable to locate many of the cases relied on by Mr. Schwartz in his brief. The cases just did not exist. Indeed, it was soon revealed that the chatbot had fabricated the cases and legal citations before feeding them to Mr. Schwartz. As described by the presiding judge, at least six of the cases Mr. Schwartz cited in his brief “appear to be bogus judicial decisions with bogus quotes and bogus internal citations.” See Mata v. Avianca, Doc. No. 22-cv-1461 (S.D.N.Y. May 4, 2023). Some of the fake opinions were cases named Martinez v. Delta Air Lines, Zicherman v. Korean Air Lines and Varghese v. China Southern Airlines
Mr. Schwartz now faces the possibility of sanctions. Mr. Schwartz submitted an affidavit to the court explaining that he used the chat bot to “supplement the legal research” for the brief. Mr. Schwartz further claimed that this was the first time he had used Chat GPT for legal research and did not know that the chatbot could provide false information. Mr. Schwarz described Chat GPT as “a source that has revealed itself to be unreliable.”
Chat GPT Doubles Down
Attached to his affidavit to the court, Mr. Schwartz submitted screenshots of his conversation with the chatbot demonstrating the attempts he made to confirm the accuracy of the legal research. Mr. Schwartz asked the bot if “Varghese” was a real case. The chat bot replied that Varghese was a real case and that the case could be found on “reputable legal databases” like Westlaw and Lexis Nexis. Chat GPT declined to admit to fabricating the cases but did offer Mr. Schwartz the following:
I apologize for any inconvenience or confusion my earlier responses may have caused.
The district court judge is still considering whether to impose sanctions against Mr. Schwartz. It’s unlikely that the chat bot’s apology for the mix up will help Mr. Schwartz avoid the sanctions.
AI Will Not Make Lawyers Obsolete
Artificial Intelligence will not—and cannot—entirely replace lawyers. The cautionary tale of Mr. Schwartz exemplifies why chatbots and other AI programs will not eliminate the need for lawyers. This is because artificial intelligence is no replacement for human experience and professional judgment. AI can and should be utilized as a tool to assist lawyers. But lawyers should not use this tool in place of their own skills and training.
Here are a few helpful tips for lawyers who do choose to utilize this tool:
- The bot may provide quick and thorough responses for simply legal inquiries. For example, if you need an explanation of trial procedures for your client, an AI program can quickly provide that routine legal description for you. The bot is not your solution for more nuanced or complex issues. The AI bot is not a lawyer and cannot interpret case law. Although the chat bot may provide intelligent sounding responses, the bot is not a qualified legal professional and is no substitute for legal advice from a qualified lawyer.
- Use AI as a starting point, not as an ending point. AI is a great resource to begin your research and fine tune search terms before taking a deep dive into Westlaw or Lexis. AI can also be a great tool to summarize and quickly synthesize complex information. But do not just cut and paste the AI’s work into your legal filings. Always personally verify the AI’s research for accuracy and to ensure that the AI program has not plagiarized its responses from another source.
- Check local court rules to verify whether judges have prohibited lawyers from using AI to prepare pleadings filed in the court. One judge in Texas, for example, now requires all lawyers and pro se litigants to file a certificate attesting that they “either that no portion of any filing will be drafted by generative artificial intelligence (such as ChatGPT, Harvey.AI, or Google Bard) or that any language drafted by generative artificial intelligence will be checked for accuracy, using print reporters or traditional legal databases, by a human being.”
- AI can simplify routine legal work and save lawyers an immense amount of time. This seems like a dream for many lawyers. But only use AI to enhance your legal work, not as a substitute for your legal work.
AI bots can mimic the work of lawyers, but the bot’s work is no substitute for a lawyer’s professional judgment. If you think that AI is going to solve your lawyer problems, keep dreaming. And don’t forget to double check those citations unless you want to find yourself in a nightmare.