By Ben Paden

Doyle & Foutty, PC *

USFN Member (IN, KY)

In May, USFN held a discussion on Artificial Intelligence (AI), and I was thrilled to serve as the moderator, which, for my part, involves not being an expert and asking questions when something is mentioned that I do not fully understand. A perfect position for me that fit like a glove! Our experts included Michael Merritt at BOK Financial, Sally Garrison with The Mortgage Law Firm, Zack Glaser with Lawyerist, and Brian Nicholas with McCalla Raymer Leibert Pierce, LLC. I extend my sincere thanks to them for their expertise and for leading interactive breakout sessions on this important topic.

I learned that Artificial Intelligence, besides being Spielberg’s worst movie (don’t @ me), is already prevalent in our industry. If you are as I was, you may have thought that there was still time to prepare for “the future” and that we would be able to slowly implement AI processes with carefully prepared and tested safety measures to protect ourselves. However, the reality is that it is already here and we may be very late to the party. So this discussion was not only eye opening, but also crucial to understanding what AI is, what it is not, how to use it properly in our jobs, and what its future use may look like.

Based on the USFNgage conversations had, let us clarify what Artificial Intelligence is and is not. AI is fundamentally technology that enables computer software to mimic human intelligence and problem-solving abilities. However, right now, we are mostly seeing what is called generative AI, machine learning and natural language processing specifically designed to create things, identify patterns, and interact with prompts. Generative AI is mostly affecting music, art, and writing at this moment. Machine learning appears in number crunching businesses, email filtering, medicine, and prepping that Amazon product for shipment once you have put it in your cart because it knows you need toilet paper and new tennis shoes just as well as you do, so why fight it? Natural language processing are your chatbots, your Clippy in Word (man, I miss Clippy), and your “Stephanie” with Visa who wants to understand why you’re calling today, but you don’t want to talk to “Stephanie,” you want a “REP-RE-SENT-ATIVE!!!”

AI is not Skynet, HAL 9000, or one of those evil robot overlords seen in a Hollywood blockbuster. At least not yet… It is more a highly sophisticated series of programs that piece together portions of the hodge-podge of humanity’s existence up until this point in order to fulfill a prompt or interaction it is given. Generative AI, for example, is solely based on material and content we have already created. It and other programs “learn” based on what we provide it. AI can draft a paper on ‘The Great Gatsby’ or generate a novel reminiscent of it, but it cannot produce wholly original creations. Moreover, it lacks true cognitive abilities – like thinking, feeling, or loving.

Having finally obtained an understanding of what AI is, I was ready to learn how it is being deployed in our workspace. We have already seen lawyers get in trouble for using generative AI in brief writing as the program “hallucinated” or made-up source material, so there are definitely misuses we need to be aware of. The consensus of the four breakout groups seemed to point to AI being used to assist in work, but requiring constant human oversight, interaction, and quality control of the outgoing product. For example, using AI to help write summarizations of large amounts of material to point an individual in the right direction for research, having AI generate a good starting point for a policy or procedure that staff then complete, or allowing AI to make simple decisions on situations to help downstream processes for employees would all be suitable uses for AI in the workplace.

You can use AI to take something you have written and rewrite the material differently, more eloquently, or more simply. I am often a terrible writer (see this article for examples), so I probably should have used AI to help me write this to make it easier and more enjoyable to read. You can have AI add an authoritative tone or, if you are too authoritative, make your writing less abrasive. AI can also be used to discover patterns or inefficiencies and reveal strengths or flaws within your own work processes. Someone can employ AI to help with audits, audit prep, or reporting. The options are almost literally limitless.

However, the danger for us still looms. Any product or procedure put in place will still be our responsibility. No one will accept the excuse of “the computer did it,” whether that is the boss, a judge, an investor, or an oversight body. It will still be our fault. Moreover, turning our thoughts outside of ourselves, it opens the door for an incredible amount of bad acting in our industry and others. Everyone has access to this technology. Fraud attempts will improve. Pro se litigant filings may be more on point and harder to dismiss or strike. Cybersecurity traffic will increase and intensify, and phishing schemes may become more successful.

Lastly, we come to the future. Here is where we envision mushroom clouds and humans living underground – a barren wasteland and bleak existences. Right? Or do we see a Star Trek-like future of blindingly fast calculations, space exploration, and elevated human existence? I am ever an optimist, so I at least hope for the latter. Generative AI, machine learning, and natural language processing programs can all be incredibly helpful, and I am not sure I necessarily see a reason to stop using them.

We do, however, need to be very careful about their oversight. The human quality control previously mentioned must continue and become even more robust. Additionally, the use of AI should be limited to non-mission-critical tasks. Money will have to be spent on staff and technology guardrails to properly utilize this kind of software. What AI is “fed” or trained on has to be carefully curated and monitored. And tested. Lots and lots of testing.

We believe we will see new roles created within companies for AI Trainers and AI Quality Control Testers, as well as some companies employing a Mathematician or Data Scientist to help control these processes. Regulations will also begin to include AI oversight, so we will have to be prepared to answer those questions. We will also probably have to train some of the regulators, as they do not always fully understand the things they regulate. Not that we have ever experienced that, right? We will have to incorporate AI oversight of our vendors, too. Just because you might not utilize AI yourself does not mean your vendors do not, and we’ll have to be aware of that as we move forward. And privacy! What to do about privacy?! An AI model can’t exactly “unlearn” something, so what do we do about data it gets hold of that we did not intend? How do we solve for an AI model that obtains PII for example?

While we could not solve for all the problems presented, I think a great job was done and the level of interest and interaction on this topic was incredible. I look forward to continuing the discussion as we move forward, both as individual companies and together as an industry. If you missed this discussion, I hope this article helped catch you up. I highly recommend being part of the next one. Be on the lookout for more USFNgage events on this and other topics in the future. It is a fantastic opportunity to come together and have an interactive discussion on a specific topic. And don’t worry, the USFNgage series is not recorded, so there’s no proof that I had no idea what I was talking about or what I was asking, which also means you too can participate freely. Thanks and we hope to see you next time!

Copyright © 2024 USFN

USFNews - July 10, 2024

* Denotes firm is a 2023 USFN Award of Excellence recipient.