ChatGPT and the Law: Like Chess, But Not as Black and White

In May 1997, Garry Kasparov – at the time, the world’s chess champion – lost a match to an IBM supercomputer known as Deep Blue. The battle of “man vs. machine” captivated the interest of not only philosophers and computer scientists but also the public in general.

More than 25 years after the showdown, which was apparently won and lost based on Kasparov’s interpretation of a glitchy move played at random, advances in machine learning have captured the general public’s attention yet again. The effects of the ChatGPT free demo have been wide-reaching, creating polarizing opinions especially within the legal community.

Our news feeds have been inundated with ChatGPT pieces for the past few months, and it’s been difficult to gain perspective. Sorting out who’s saying what, and why, has provided a bit of clarity. And to some of us, the “celebrity” status the “bot” has reached seems a lot like the hype for Deep Blue.  Public awareness may have reached a critical mass in spring 2023 with President Biden trying out ChatGPT as part of a White House agenda related to generative AI.

To speculate a bit, it seems to us that the chess community of the 1990s felt the same pessimism when a computer program defeated a grandmaster that some attorneys feel at the perceived threat of ChatGP. After all, what’s the point in developing a word-class skillset only to be bested by a microchip?  OpenAI’s generative AI, in only its 4th iteration, can already ace the LSAT and the multistate bar exam.

Other attorneys marvel at the software’s capabilities at increasing access to justice – quality advice is available to all, finally!  Or: my small firm can compete with the big firms on breadth and depth of service!  Flashing back to the 90s again, this was the amateur chess player feeling capable of beating Kasparov.

It doesn’t take too much advanced training to predict our argument here: the answer is somewhere in the middle. Continuing the chess analogy, players today routinely use computers to train. And this tool (i.e., widespread access to swaths of game data and strategy) has not only broken down barriers to entry for chess newcomers but has also raised the top level of the best of the best.

The ChatGPT “bot” has access to millions of data sources. Its answers aren’t always “right” (and with the millions of savvy authors it’s “read”, the outputs often look spot-on even though they’re super vague), so knowing when the computer doesn’t deliver an efficient answer is the trick.

Sure, some attorneys will use the new technology to dabble. And yes, in the same way ChatGPT can do your junior high school student’s homework, it can also do legal research for lawyers who are strapped for time. But given our practical experience, we can forecast that an overwhelming reliance on technology to do the heavy lifting will likely bring about inefficiencies overall.

As attorneys and counselors, we owe it to clients to understand the technology and respond rather than react. After all, when the ride-sharing software tells the driver to take Lower Wacker, the backtracking required for an unfamiliar navigator will wipe out any time the computer thought it’d save you.

This document is intended for informational purposes only and is not legal advice or a substitute for consultation with a licensed legal professional in a particular case or circumstance.

Jonathan S. Safron | Construction Law, Mechanic’s Liens, and Creditors’ Rights

Jonathan is a litigation attorney at Carlson Dash. His practice focuses on construction, mechanic’s liens, and creditors’ rights. If you need assistance with a related matter, contact Jonathan.