Why does it seem like every tech blunder story involves an attorney? First, the cat filter Zoom call fiasco had everyone in stitches, and now it’s the ChatGPT attorney. The New York attorney relied on ChatGPT, an artificial intelligence chatbot, to do his work for him. I wasn’t sure if I wanted to write about this latest blunder but the details were just too juicy to pass up.

The New York attorney was the plaintiff’s attorney in a case involving a passenger injured by a drink cart on an airline flight. The attorney was working on a response to the airline who asked the judge to dismiss the passenger’s lawsuit because the statute of limitations (SOL) had passed. Every type of lawsuit has a time limit in which it can be brought. The SOL is a hard and fast rule; it’s almost always game over, do not pass go, do not collect $100. 

The attorney asked ChatGPT to write a brief arguing why his client should be able to move forward despite their lateness. Known for refining and steering a conversation, that’s exactly what ChatGPT did. It spit out a legal argument complete with case law. The problem was the cases didn’t even exist! The courtroom was packed during the two-hour hearing, during which the attorney struggled to explain himself. His nephew or some young whippersnapper had told him about ChatGPT. Law students were in attendance. Multiple news outlets covered the story. I even saw my old law school professor of Ethics and Professional Responsibility, Rebecca Roiphe, quoted in an article about the dangers of using ChatGPT. 

I understand the temptation to use AI at work. Even Lexis Nexis has been sending emails to attorneys promoting their new AI products (for a fee). I recently used ChatGPT (for free) to draft a demand letter as a test. It was impressive. Next, I asked it to explain how to set up a parent LLC and subsidiary LLCs in the context of real estate. The answer was spot on and fast. This leads me to my first question, which I am dying to know if the judge asked, how much did the ChatGPT attorney bill his client for the brief? Obviously, the attorney didn’t look up the “cases” in trusted legal databases, like Westlaw or Lexis Nexis, or he would have realized something was off. 

Only time will tell how AI will impact different industries. I just read about Meta’s Llama 2, the latest from our good pal Meta (Facebook) Chairman and CEO Mark Zuckerberg. Instead of keeping the coding behind this language model, a competitor of ChatGPT, secret, he’s decided to make the model open source. Llama 2 has been trained on 40 times more data than its predecessor. Zuckerberg believes that open source will lead to rapid progress, the elimination of bugs and healthy competition. The flip side is that safety may be compromised. Malicious actors could misuse AI to generate disinformation and spam. Admittedly it is harder to regulate something that is already in large scale use. Google and Microsoft have limited access to their latest AI because of safety concerns and competition. Is AI a Pandora’s Box? Will we all be influenced by A.I.? Only time will tell.