[Editor’s Note: EDRM is proud to publish the Hon. Ralph Artigliere’s (ret.) advocacy and analysis. The opinions and positions are Judge Artigliere’s (ret.) June 7, 2024 © Ralph Artigliere.]
Judges and lawyers received a significant gift last week when 11th Circuit Judge Kevin C. Newsom penned a concurring opinion in a seemingly mundane insurance case involving a backyard in-ground trampoline. The concurring opinion, however, transcends its humble context, presenting a visionary outlook on a pivotal issue: the integration of advanced technology, particularly AI, into judicial workflows. As we navigate the current concerns surrounding AI—such as safety, hallucinations, and bias—this opinion prompts us to consider how we can harness machine intelligence to enhance judicial efficiency and efficacy.
Judge Newsom’s thoughtful opinion is a must-read for litigators and judges alike, as it addresses a critical question: what role will technology play in judicial workflows and decision-making once we overcome concerns about safety, hallucinations, and bias?
The Hon. Judge Ralph Artigliere (ret.)
The opinion is found on LexisNexis here: Snell v. United Specialty Ins. Co., 2024 U.S. App. LEXIS 12733 *; _ F.4th _ (11th Cir., 05/28/24). Ralph Losey has also provided the entire opinion and footnotes and illustrations along with his always enlightening insights here. The concurring opinion, which Judge Newsom himself said was unusual—even for him— and “Just [his] two cents,” is actually worth a million bucks to the rest of us.
Judge Newsom’s thoughtful opinion is a must-read for litigators and judges alike, as it addresses a critical question: what role will technology play in judicial workflows and decision-making once we overcome concerns about safety, hallucinations, and bias? The discussion has begun, and Judge Newsom’s opinion provides a valuable framework for exploring how AI can enhance judicial efficiency and effectiveness while maintaining rigorous standards of fairness and propriety.
Machines and the Work of Judges
Just as technology has revolutionized business, government, and personal lives, machines will inevitably influence the work of judges and lawyers. AI, in particular, has become too powerful and sophisticated to ignore. While it is still early and challenges remain, AI products are rapidly improving in safety and effectiveness. Legal professionals are still learning how to use generative AI properly and safely. Judges and lawyers can and must adapt and incorporate these advancements to enhance their work, leveraging the brainpower and efficiencies of machine intelligence.
It is genius because the vehicle was perfect: a specific narrow issue in a concurring opinion so that the real message could be laid out in a transparent, thorough, and balanced manner without carrying the weight or limitations of judicial precedent.
The Hon. Judge Ralph Artigliere (ret.)
Meanwhile, through the genius and courage of a federal appellate judge, the cards were laid on the table: why wouldn’t judges use AI to assist them in reaching the truth and a just result, provided it can be done safely and within the rules? I say genius, because it is so cool for an appellate judge to use a concurring opinion in a small, routine case to start a dialogue on one of the most intriguing issues of our time. It is genius because the vehicle was perfect: a specific narrow issue in a concurring opinion so that the real message could be laid out in a transparent, thorough, and balanced manner without carrying the weight or limitations of judicial precedent.
The opinion is courageous, as Judge Newsom’s views are bound to be controversial, particularly regarding the consultation of resources outside the record. However, his advocacy for using technology to find the common meaning of “landscaping” demonstrates the potential for AI to assist in judicial decision-making without compromising judicial integrity.
But before I get ahead of myself, let’s look at the opinion and see what others have said about it already.
The Snell Concurring Opinion
In this insurance coverage and bad faith case, the parties litigated the matter as an “ordinary meaning” case, arguing over whether Snell’s installation of an in-ground trampoline, retaining wall, and a decorative wooden “cap” fit within the common understanding of the term “landscaping” in Snell’s insurance policy with United Specialty Insurance Company. Because Snell’s claim was defeated on other grounds based on Alabama law and the underlying facts, a determination on the issue of whether the in ground trampoline construction was landscaping was not needed in the 11th Circuit’s decision. But, because that “off ramp” [Judge Newsom’s words] was not obvious to the Court from the outset, Judge Newsom spent “hours and hours” laboring over whether this backyard construction was indeed landscaping in common parlance.
After recognizing that a foray into AI actually looked like a fruitful path, and after the case was decided on other grounds, Judge Newsom directed his efforts to a discussion of the pros and cons of judges using AI in the fashion he would have used it if needed in his case.
The Hon. Judge Ralph Artigliere (ret.)
Judge Newsom offers an unvarnished glimpse into his judicial reasoning, particularly his struggle to define “landscaping” in the context of an insurance policy. His contemplation of the potential advantages of using AI tools, such as ChatGPT and other Large Language Models, underscores the transformative potential these technologies hold. Judge Newsom illustrates how AI could provide efficient and accurate interpretations of legal terms, thereby expediting the judicial process and enhancing the precision of judicial decisions.
After recognizing that a foray into AI actually looked like a fruitful path, and after the case was decided on other grounds, Judge Newsom directed his efforts to a discussion of the pros and cons of judges using AI in the fashion he would have used it if needed in his case. Judge Newsom does a good job laying out the advantages and potential pitfalls.
Judge Newsome points out and discusses in detail numerous advantages and potential pitfalls against the backdrop of judicial decision making in cases searching for the ordinary meaning of contractual terms. The advantages of Large Language Models in that context are: LLMs train on ordinary-language inputs; LLM’s can “understand” context; LLM’s are readily accessible; LLM research can be relatively transparent; and LLMs hold advantages over other empirical interpretive methods. The potential problems are: LLMs can “hallucinate;” LLMs do not capture offline speech, and thus might not fully account for usage of underrepresented populations; lawyers, judges, and would-be litigants might try to manipulate LLMs; and reliance on LLMs will lead us into dystopia. 2024 U.S. App. LEXIS 12733*32-53. Each of the pros and cons are discussed with care and due consideration. The opinion cautions that LLM tools should be employed with caution and humility, quoting and agreeing with Justice Roberts’ Year-end Report (discussed below).
The opinion concludes:
But—and this is my bottom line—I think that LLMs have promise. At the very least, it no longer strikes me as ridiculous to think that an LLM like ChatGPT might have something useful to say about the common, everyday meaning of the words and phrases used in legal texts.
Just my two cents.
11th Circuit Judge Kevin C. Newsom, Snell v. United Specialty Ins. Co., 2024 U.S. App. LEXIS 12733 *53; _ F.4th _ (11th Cir., 05/28/24).
Indeed, Judge Newsom. Thank you for saying it in an official capacity.
Food for Thought and Conversation Among Lawyers and Judges
The concurring opinion displays how judges think and how generative AI might fit in helping them to reach just decisions more efficiently and perhaps better than they would using traditional resources. Using generative AI tools will not be an easy path, especially with platforms in their current nascent stage. Large Language Models and platforms that incorporate them in legal tasks require understanding of their limitations and pitfalls as well as a thorough understanding of precisely when they can be used and how to operate them properly with good prompts and a careful review of output. But Judge Newsom’s opinion covers all that and more.
There have already been rumblings over the opinion. Wilkins, S., 11th Circuit Judge Uses ChatGPT in Deciding Appeal, Encourages Others to Consider It, ALM Law.com (June 4, 2024). In that piece, the title is regrettably misleading and there is a statement that suggests that Judge Newsom “laid out his use of the generative AI chatbot to help inform his analysis of a key issue in an insurance appeal.” As noted above, the research on the term “landscaping” actually turned out to be unrelated to the basis of the majority decision.
Judge Newsom’s opinion raises issues of propriety, fairness, and the sanctity of consulting external resources without party input.
The Hon. Judge Ralph Artigliere (ret.)
Judge Newsom’s opinion raises issues of propriety, fairness, and the sanctity of consulting external resources without party input. These concerns, echoed by Judge Scott Schlegel of the Louisiana State Court, highlight the need for stringent guidelines and ethical standards. Schlegel, S., The 11th Circuit’s Experiment with AI: Balancing Innovation and Judicial Integrity, Substack (June 5, 2024). Judge Schlegel, while recognizing the innovative potential, emphasizes the balance between embracing technology and maintaining judicial integrity. This balanced approach should be a touchstone for measuring steps into new technology.
Issues such as boundaries of judicial research are precisely why Judge Newsom’s willingness to write on this topic to be ground-breaking and a great kickoff for discussion. Judges have traditionally referred to dictionaries to inform them on common understanding of terms in contracts. Judge Newsome’s opinion lays out exactly why AI, carefully used, is a broader and more current source of common understanding of terms than a stack of dusty dictionaries. I agree with Judge Newsom on that point. But my opinion is just that… one opinion. We all must stay tuned and engage the dialogue.
Why the Issue of Scope and Means of Inquiry by Judges Is Important
Judges want to make just decisions, and they prefer to have all the information needed to do so. But judges need to stay between the lines in their research. The parties are entitled to decisions that are based on the facts in the record. Legal research is one thing. Fact research by a judge without the opportunity of the parties to review and respond to sources is quite another. Judges must remain neutral between the parties, which can be threatened by engaging in research beyond the record.
Trial lawyers do not want to be out of the loop and prefer to drive the train. An old saw when I was a trial lawyer went something like this: “Your honor, I do not mind if you try my case for me. But if you do so, please win it.”
The Hon. Judge Ralph Artigliere (ret.)
Trial lawyers do not want to be out of the loop and prefer to drive the train. An old saw when I was a trial lawyer went something like this: “Your honor, I do not mind if you try my case for me. But if you do so, please win it.”
To the point of Judge Newsom’s opinion: Can emerging AI tools help judges get the information they need to decide cases correctly and efficiently? Use of generative AI by lawyers and pro se litigants could be a tremendous boon for access to court and communication between the parties and the judge. On the judicial side, AI will undoubtedly help the judge and staff to marshal information, analyze and organize it, and prepare orders and decisions. Development of the tools is furious but remains nascent. The issue is still open on whether generative AI is ready for a primary role in research, and we know human oversight and validation will always be with us. To be prepared, judges need to explore and understand the pros, cons, cautions and stop signs along the way. That is the wisdom of Judge Newsom’s opinion.
Thank You Judge Newsome and Justice Roberts
I find it refreshing that judges at the highest level are discussing the transition to AI. There are other examples. Like Judge Newsom, others are thinking ahead and considering the potential value of properly employed advanced technology.
Chief Justice John Roberts made the rise of AI a feature of his Year-end Report on the Federal Judiciary released Dec. 31, when he said judicial work “will be significantly affected by AI.” Justice Roberts both recognizes the inevitable tide of AI use in courts and hits the same refrain we see in Judge Newsom’s cautions and Judge Schlegel’s expressed concerns. But the upside and the challenge to evolve is definitely expressed in the Report:
Rule 1 of the Federal Rules of Civil Procedure directs the parties and the courts to seek the “just, speedy, and inexpensive” resolution of cases. Many AI applications indisputably assist the judicial system in advancing those goals. As AI evolves, courts will need to consider its proper uses in litigation.
Chief Justice John Roberts, Year-end Report on the Federal Judiciary (December 31, 2023).
Another example comes from Judge John Bush of the US Court of Appeals for the Sixth Circuit who predicted AI could ease historic linguistics research in remarks made during Speaking at a Federalist Society event last March at the University of Chicago Law School. Monyak, S., AI to Make Originalist Historical Analysis Easier, US Judge Says, Bloomberglaw (Apr. 1, 2024). Judge Bush said the originalist approach to interpreting text “is consistent with, and indeed will thrive” with advancements in AI and that AI could help in the laborious task of examining word usage in historical context.
CONCLUSION
Research by judges may be the most controversial application of AI. Significant challenges lie ahead, but generative AI products will continue to evolve and improve. Justice Roberts is right: addressing these challenges directly will pave the way for a more efficient and effective judicial system. Judge Newsom’s groundbreaking opinion is more than a “shot across the bow;” it is a clarion call for the legal community to embrace the future of AI thoughtfully and courageously in judicial decision-making.
June 7, 2024 © Ralph Artigliere. ALL RIGHTS RESERVED (Published on edrm.net with permission.)
NOTE: Generative AI products were used to help review this article per EDRM GAI and LLM Policy.
