Over the past few days, I have experimented with an artificial intelligence product from a well-known legal research provider. I ran the product through a number of common litigation tasks. Full disclosure: I’ve been casually reading about AI in the legal field since 2019, but I’m not an expert. In fact, I am an AI novice in every sense of the word. But, like many of you, I have recently become very curious about AI and how it might change the practice of law. So, here’s what I did with my 7-day free trial from the AI provider:
- I asked the AI product a series of questions that I already knew the answers to. The AI product provided very well-written responses. Perhaps most important, the answers were correct. I ran the exact same questions through Lexis’s natural language search and quickly found the same law. I suspect quality and reliability would decrease with the AI product as the legal questions become more nuanced, complex, or fact-driven.
VERDICT: The legal research function of the AI product worked well, but a well-crafted Lexis search ended with the same result. At this time, I don’t see any real practical benefit from this AI product over my existing legal research tools.
2. I uploaded several depositions from old cases and asked the AI product to summarize the depositions. One of the depositions was a two-day-long party deposition in a construction defect case. The deposition summary for the construction defect case contained entries like: “Deponent discusses _____________” and “Deponent is asked about ___________” This type of summary is only useful to obtain a very high-level understanding of the deposition. But, to be fair, the AI summary contained some relevant substantive content that was helpful in understanding the deponent’s testimony. I happened to have a detailed line index summary prepared by a paralegal for the very same deposition. The deposition line index prepared by the paralegal was well done, relevant, and specific. I do not routinely have line indexes compiled, but that particular deposition was important. And, with the client’s permission, we invested a day and a half of paralegal time to get it done. The paralegal’s (human) effort had a meaningful impact on the outcome of the case. I don’t believe the AI-generated deposition summary would have had the same impact.
I also asked the AI product to summarize a plaintiff’s deposition in a simple car wreck case. The AI output was better for this deposition. But, the summary still contained entries like “Deponent discusses _______”.
After redacting personal identifying information from a set of medical records totaling over 600 pages, I asked the AI product to summarize the records. The AI-generated summary contained a multitude of repeated information. It was almost as if the AI product summarized each page of the medical record in a vacuum rather than looking at the record as a whole, or in context. For example, the plaintiff’s date of admission was mentioned at least 20 times throughout the summary. The plaintiff’s diagnosis and the chief complaint were listed many times as well. Perhaps if a paralegal sifted through the records, removed irrelevant documents and redundant documents, and then asked the AI product to summarize, the AI output might be better. Coincidentally, my paralegal prepared a medical record summary for the same set of records. The summary is well done. It’s not repetitive. It highlights the relevant information and takes a deep dive only where appropriate.
VERDICT: The summary function was not helpful. I’m going to stick with my in-deposition summaries, PDF word searches, and when appropriate, paralegal-prepared deposition line-indexes.
3. Lastly, I tried the ask a question function. The function allows you to upload a set of documents and ask specific questions about those documents. I uploaded redacted medical records from the automobile accident, depositions in the same case, and, the accident report. I asked the following questions: “How did the accident occur?” “Who witnessed the accident?” “Were alcohol or drugs a factor in the accident”, and “What were [name of plaintiff]’s injuries?” The results were surprisingly good—in light of the poor deposition and medical record summaries. In fact, with some light editing for content and context, the output is good enough to send directly to a client, and good enough to help with an early case evaluation. I also uploaded a dozen depositions in a construction defect case. I asked the AI product a series of questions that I already knew the answer to. The output was good, but not great. The results were technically accurate but superficial. The results lacked context and were void of analytical thought.
VERDICT: Overall, this function provided meaningful and accurate responses to specific questions. However, I suspect as the complexity and size of a case increase, the usefulness and reliability of the responses will decrease. Artificial intelligence is one of several cutting-edge technologies my Firm and I are evaluating as part of our commitment to provide high-quality and cost-effective legal services to our clients. However, AI remains in the infant stage of development.
Based on my limited review, the AI product has only limited practical uses at present. I have no doubt that the technology will improve. Consequently, we will continue to monitor and evaluate newer iterations of AI products. I am committed to high-quality cost-effective legal services. So, I don’t see myself leveraging AI products (summaries, research assistance, document drafting) unless and until the products constitute a real value proposition to my clients.
M. Jansen Voss is a partner in the Birmingham office and has developed a diverse defense litigation and appellate practice in both state and federal courts in Alabama and Mississippi. He represents a wide range of businesses, governmental entities, and individuals in complex personal injury and wrongful death lawsuits, as well as business disputes and breach of contract matters.
About Christian & Small
Christian & Small LLP represents a diverse clientele throughout Alabama, the Southeast, and the nation with clients ranging from individuals and closely-held businesses to Fortune 500 corporations. By matching highly experienced lawyers with specific client needs, Christian & Small develops innovative, effective, and efficient solutions for clients. With offices in Birmingham, metro-Jackson, Mississippi, and the Alabama Gulf Coast, Christian & Small focuses on the areas of litigation and business, is a member of the International Society of Primerus Law Firms, and is the only Alabama-based member firm in the Leadership Council on Legal Diversity. Our corporate social responsibility program is focused on education, and diversity is one of Christian & Small’s core values.
No representation is made that the quality of legal services to be performed is greater than the quality of legal services performed by other lawyers.