Martin is a lawyer who has closed many residential real estate transactions. Back when he was starting out, Martin read the relevant statutes and regulations and prepared all the documents himself. For each transaction, Martin drafted a Statement of Adjustments, adding numbers in his head. Eventually he used a calculator, and then a software program. Then, he hired a law clerk and taught her how to do it. Should the situation arise where Martin needs to prepare a Statement of Adjustments manually, he can easily do so, because he knows everything that there is to know about them. When a dispute arises as to the accuracy of a Statement of Adjustments, he can resolve it.
Eric has not been around nearly as long as Martin. When Eric started his law practice, he did not read the statute or the regulations for his practice area, but he did google a few articles and read them. He bought some software and hired an experienced clerk. Eric has also closed quite a few residential real estate transactions. Most of the work on his files is done by his law clerk, using the software program.
The other day, Eric had to prepare a Statement of Adjustments for an asset sale transaction. Not having a software program do it for him, he did not have a clue as to where to start. You see, Eric never took the time to understand how a Statement of Adjustments is put together. If a dispute concerning a Statement of Adjustments cannot be fixed by asking his clerk to check the data that was entered into the system, Eric will have a problem resolving it. He simply does not have a good conceptual grasp of the transactions that he is responsible for.
Lawyers who learned how to do things themselves will naturally have a deeper understanding of what they are doing than those who have only relied on staff and software to produce their work product.
Eric, and others like him, who I have observed over the last few years, concern me.
Now, along comes AI, and the problem is increasing exponentially. Some young lawyers are producing documents and developing their legal opinions by providing prompts to AI, having never done the work themselves. AI might be smart, but it is not yet smart enough to figure out which questions it should be answering.
We have all heard the expression, “You ask a stupid question, you get a stupid answer.” I would rephrase that to the less pithy expression, “You ask AI a stupid question; then stupidly rely on the answer because you don’t know any better; you get a negligence claim.”
I am not feeling good about the future of the legal profession. I could do some research on the topic, but I think that I will just check in with Chat GPT and see what it thinks.
This article was originally published by Law360 Canada, part of LexisNexis Canada Inc.