If you own a law firm in 2023, you need to discuss A.I. We do, so here goes: A.I. is the hottest trending news across the internet. Not A.I. as in what's probably the most overlooked Spielberg movie but ChatGPT A.I., the latest technology that “will revolutionize a lot of industries,” law among them. So it's claimed.
It's also claimed that the high school essay is dead; colleges are about to be buried in plagiarism; copywriters, screenwriters, reporters, and editors are about to be out of jobs; and there'll be a seismic shift in the insurance and medical industries, among many others.
You can’t open up a newsfeed without seeing something about AI devastating entire professions, undoubtedly well on 'its' way toward full consciousness, Skynet, the nuclear destruction of a Los Angeles playground, and the rise of Sarah Connor.
We've read reams of stories, caught John Oliver’s AI story a few weeks ago, and extensively tried and tested ChatGPT. We are here to report that . . . it may become a valuable tool for our everyday practice. That's it.
An other tool. Its value will lie in its ability to take over some mundane, repetitive tasks, thereby giving everyone at Hopkins Centrich more time with clients.
In his review of Spielberg's AI, Roger Ebert wrote, "We are expert at projecting human emotions into non-human subjects, from animals to clouds to computer games, but the emotions reside only in our minds . . . A. I. is not about humans at all. It is about the dilemma of artificial intelligence. A thinking machine cannot think. All it can do is run programs that may be sophisticated enough for it to fool us by seeming to think. A computer that passes the Turing Test is not thinking. All it is doing is passing the Turing Test."
Emotions. They're one reason why AI and our practice won't be compatible anytime soon.
Another is imagination. It doesn’t put things together or make connection ‘leaps.’ We tried to reverse engineer an earlier newsletter article – the one about the Titanic and Non-Compete Agreements. (Look below for a link if you missed it). ChatGPT knew a lot, instantly, about the Titanic, the Senate ship safety hearings, the subsequent ‘lifeboat law,’ and the S.S. Eastland but could not make the links (or any) between them. It took at least three-times longer to take ChatGPT through it than it did to write it in the first place.
We envision using ChatGPT for deeper and much faster searches but that's about it.
Because what we do centers – always – around listening. Then analyzing, then designing, talking, explaining, re-designing, and drafting to make it all match what our clients need now and in the future. That takes knowledge and experience, it's not about patterns or flowchart decision making.
This is what makes business law uniquely challenging, uniquely rewarding, and uniquely unfit for 'do-it-yourself' solutions, such as the use of AI for anything beyond research. It absolutely requires a knowledgeable, experienced practitioner because AI has a problem: when in doubt of an answer it will fake it. Then it runs with the facts it made up leading to a long string of impeccable logic based on the initial false premise.
This is not a 'glitch,' it's part of the design. It's a way for the AI to learn but, of course, it only works if the person typing on the other end knows ChatGPT is wrong. Otherwise, it will go on to create the mother of all informational rabbit holes.
In the hands of someone not conversant in every aspect of business law, that’s a very, very deep rabbit hole.