Where the Rubber Meets the Road: The Intersection of Generative AI and Legal Data Intelligence

The Legal Data Intelligence (LDI) model was created to equip legal professionals with a holistic framework, shared vocabulary, and common set of best practices for a wide variety of use cases. Armed with a new model with which they can address the data challenges that keep lawyers up at night, LDI practitioners have a unique opportunity to unlock the benefits of AI.

With 2024 being an inflection point for large language models and 2025 projected to be a seminal year for autonomous AI agents, AI models are showing exponential improvement in their performance and level of sophistication. Legal Data Intelligence practitioners—with their experience overseeing massive data projects, building defensible workflows, and working with artificial intelligence to find novel solutions to complicated data problems—will play an instrumental role in helping their industry peers stay abreast of, and adapt to, the technology shifts on the horizon.

Over the course of the last year, founding members and proponents of the LDI model have shared their insights on how the emergence of generative AI will change the way the industry works, along with the new opportunities and challenges that these changes will create.

Here are some of the observations we’ve drawn from those discussions.

A Conversational Approach to Interrogating Data

Founding member Bobby Malhotra (Litigation Partner, Winston & Strawn) believes that large language models (LLMs) are altering how legal professionals query data to find answers. In the past, interrogating data required the use of Boolean searches—a rudimentary process that, according to Malhotra, was prone to delivering results that were both under-inclusive and over-inclusive.

“As an attorney, you mainly relied on your technical support folks to gather and spoon-feed you the underlying information, which you then had to sift through,” said Malhotra on a recent podcast episode.

But much like the use of LLM interfaces such as ChatGPT has made users accustomed to asking questions to the AI model in conversational language, legal professionals could use AI models to interrogate data by asking questions: “You can ask the model in human language what you want to see and what you want to surface. So instead of these complex Boolean searches and separate proximity operators, you're using a conversational tone to get to the data you need, which I think is far superior and has changed the game.”

Embracing Ethical Duties and Guardrails in the Age of Algorithms

According to Daniel Gold (Principal, BDO USA), Legal Data Intelligence gives legal professionals a network where they can learn about and share their lived experience of using generative AI technologies in legal matters.

“Attorneys have an ethical responsibility to learn about the risks of new technologies like generative AI,” said Gold on a recent podcast episode.

Gold cited Rule 1.1, Comment 8 of the Model Rules of Professional Conduct by the American Bar Association, which was amended in 2012 to include knowledge of relevant technology in the attorney’s duty of competence.

Gold noted that the popularization of LLMs is quite unlike the adoption curve with previous technology paradigms: “When we talk about generative AI, you’re talking about very, very fast adoption. Well, why is this happening? It’s because we are talking about generative AI at the dinner table, your kids are talking about it, your spouse is talking about it. It’s readily accessible and it’s free. It’s not like 14 years ago when we were talking about technology-assisted review (TAR).”

In such a fast-changing landscape, Gold argues, it’s incumbent upon legal professionals to take precautions aimed at preventing the inadvertent disclosure of confidential information that is prohibited under Rule 1.6 [C].

“When everyone’s talking about it [generative AI] and it’s so readily available, there is more danger of potentially breaching Rule 1.6 [C] by putting information into a generative AI tool.”

Prompt Engineering: A New and Vaunted Legal Skill

Over the course of the last two years, the pace of technological change has been overwhelming, with our definition of what’s normal and what’s possible changing every few months. But if there’s one thing that will predictably become more relevant with time, it is the importance of prompt engineering, according to Legal Data Intelligence founding member Adam Rouse (Senior Counsel and Director of eDiscovery Operations, Walgreens).

“I truly believe that within three to five years, there will be job descriptions for prompt engineering. Because the AI has to be told very, very specifically, in a very creative way, how to do what we want it to do. So we're going to need people who are experts at prompting the machine to do what we want it to do,” said Rouse on an episode of the Legal Data Intelligence podcast.

Rouse believes the technological shifts will inevitably impact how and whom legal departments and law firms will hire in the near future: “I think there will be a different kind of job-hunting experience. The people who are comfortable with AI and know how to operate it at a basic level will have an edge over those who don’t in the job market.”

Principles Over Specifics: A New Approach to Navigating the AI Era

LDI founding member Briordy Meyers believes that the increasing adoption of generative AI in legal workflows will require a more flexible, reflective, and experimentative approach to solving legal data challenges.

“You can’t be too specific when things are changing all the time. You won’t have a perfectly baked model for a static use case,” he said in an interview with podcast host David Cowen.

To illustrate his point, Meyers drew a contrast between TAR and generative AI, noting how the latter is nearly impossible to deconstruct: “You can’t try to explain the black box the way that we used to try to with TAR, because you have algorithms that are creating their own algorithms that are creating their own algorithms and so on."

The path forward, he believes, lies in taking a principles-based approach akin to the EU AI Act.

“Principles can be abstract and that can be very jarring for someone who likes to work hard, get to a place, and settle down. But if you embrace the dynamic and changing world that we are entering into, it can be a very exciting time because your skillset will organically allow you to get to a place where you can bring genuine value. There will be many opportunities to blaze your own trail,” Meyers concluded.

For a deeper look at how Legal Data Intelligence leaders have leveraged AI to address data challenges across various use cases, check out this recent blog post featuring detailed LDI case studies.