Artificial intelligence war stories - by Thomas Kearns

The proliferating artificial intelligence products for law firms show promise of improving the quality of legal services and making lawyers more efficient, but as with any developing technology, the results are mixed, with some results being not worth the effort and others simply wrong. A few examples may help guide the expectations of real estate professionals and their counsel. Of course, whenever a lawyer with 40+ years of experience writes about new tools, there will be some in the audience who will see inherent resistance, but this senior lawyer is an AI optimist and a strong proponent of the ultimate benefits to the legal practice.

First example: Over the years, I have developed checklists to use when handling various matters to remind me of issues I or my clients have experienced. One of those checklists covered items to review in representing minority members about to enter into limited liability company agreements. The checklist serves as sort of a wish list for minority investors. LLCs typically vary widely from heavily skewed in favor of the managing member, to balanced, to simply poorly drafted.

When a client recently asked me to review a draft LLC agreement for a project he was thinking about investing in, I was able to load my checklist and the proposed draft agreement into one of our AI tools with a prompt, asking the tool to compare the draft to my checklist and prepare a memo listing areas where the draft fell short. The results were impressive as a first draft of a memo. It needed work and my judgment as to what edits to ask for, but it was a very helpful start.

A less successful attempt involved prompting the AI tool to review a draft sale contract for buyer side issues. The result was a mess — minor issues were given greater weight. The tool had no clue as to New York customs. It missed typical compromises. The attempt wasted time instead of helping.

We are working on a project where we take a large group of precedents, memos and other materials covering a specific typical agreement used in commercial real estate transactions and upload them into specialized database within one of the tools to permit creation of drafts based on the precedents and answer queries as to common ways of drafting certain clauses. The project shows promise but to do it right requires a lot of review of prior work product and editing.  Creation of the database is time consuming, but we expect that hard work will pay off for future client service on that type of agreement. 

Lawyers will continue to work with these tools and develop ways to have them improve client service, but they are not magic, and their effectiveness will be greatly enhanced by the lawyer’s own input of checklists, samples, and detailed prompts. Experienced lawyers may use AI to double down on their areas of expertise. If a lawyer has a large database of a certain kind of transaction documents that expertise will be magnified by AI. One concern I hear often — that AI will eliminate the need for junior lawyers — will not happen, at least at midsize or smaller law firms. All of the examples above were all teed up for my review by less experienced lawyers. It turns out I have more ideas on how AI can help us in our practice than I have time to implement by myself!

Thomas Kearns is a partner with Olshan Frome Wolosky LLP, Manhattan, N.Y.