Are advanced legal research tools a threat to lawyers? Like many emerging technologies, artificial intelligence-enabled research makes us question some of our oldest assumptions. How does a model built on exclusive access to databases and research techniques survive machines doing those things better and faster? Rather than compete with the machines, you should think of these tools as allies. Here’s why.
…On November 6th, 1983, a hallway built beneath Hawkins, Indiana flickered under dying lights.
The scientist couldn’t escape. Waiting for an elevator door that wouldn’t close fast enough, he saw something in the shapeless black. He heard it growl. And then it ate him.
That first harrowing scene of Netflix’s Stranger Things relied heavily on artificial intelligence (AI). The date, the look, the location, the monster-driven plot, the marketing that brought it to your attention—all were choices informed by applied data science.
Still, human artists created Stranger Things. They just happened to know you’d love it.
This relationship between tool and artist reveals a pattern that you can apply. When you bring advanced data to your work, your artistry improves.
Data and Artistry Combined
Netflix’s foray into 80s-style strange began with you. Well, you and everyone else who watches Netflix.
With each button click, preference setting, and late-night binge, you told Netflix what you like. The company’s advanced data tools tracked those clicks and binge sessions, looking for patterns. And they found them.
Netflix’s viewers liked shows set in the 80s. They liked ensemble casts of heroic kids dealing with middle school drama. They liked monsters and Paul Reiser and the Cold War. All essential elements, but not a TV show.
A TV show required artists who’d work within the constraints the data defined. This reveals a better model for your modern law firm.
Traditional Sources of Law
My understanding of the model firm began in law school. On the first day, I went to the library. I found aisles and aisles of books. During my legal education, I learned that clients wanted insights, but navigating the books came first. I signed up for a couple of free accounts with traditional legal research companies and learned terms like “Boolean” and “connectors.” They helped me navigate the books, but the patterns were hard to find.
Years later, I rented a room in a law office. Walking in, I found another shelf full of matching books. I sat at my desk and logged into an expensive legal research program. The program made the books more accessible, but it didn’t make discovering patterns much easier.
Then I created a website for my new firm. Searching the internet for lawyerly images, I found pictures of more books—lawyers in front of shelves, leaning forward and looking stern. The books seemed less a way to find insight-inspiring patterns and more a way to say I’m a lawyer. Call me.
The exclusivity of the bookshelf became a stand-in for actual client value. That couldn’t have been right.
A Different Approach to Legal Data
More recently, I walked into online research company, Casetext’s, office in San Francisco. I saw another wall of books, an odd fixture for an innovator trying to eliminate the need for those giant reporters. Laughing at the irony, I asked CEO Jake Heller why the company had that collection.
“They make for a funny wall to hide the bathroom,” Jake said. “The door is right behind the bookshelf.”
He then pointed to a sign on another wall. “Free the law,” he read, words scribbled in neon lights. “That’s what we really do.”
The mission startled me. Jake explained that making law freer made patterns and insights easier to find, but I resisted. Moving the law from library shelves to every home desktop, then adding tools that analyze the data at record speeds? That felt like a threat.
As a solo attorney, I’d never tried to make the law free. I benefited from owning shelves of books that I’d never read and that no one else could interpret. The books were a symbol of exclusivity. That was the model.
My bookshelves sent a signal to prospective clients: “I alone have the words, so you need me.” The words in those books gave my business a foundation proven profitable for over a hundred years.
But things have changed. The internet has turned words into a commodity. Consumer respect for the profession has eroded and their demands for accuracy and creativity have increased, even as they push back on high research fees. They want more for less. Lawyers will need a new business foundation.
So what will we sell if not exclusive access to the sources of law? What’s our foundation if not books full of words?
The Promise of Advanced Legal Research
Valerie McConnell, attorney product specialist at Casetext, showed me what more advanced legal research tools can do. She found a document on her computer’s desktop, dragged it onto a portion of Casetext’s website marked “CARA,” and asked what issue I wanted to focus on.
She didn’t need an unnatural string of phrases, or slashes and ampersands, or a defined number of words between two terms. She just wanted a general concept that might be on my mind.
I picked a random issue out of the air. Valerie plugged it in, waited a few seconds, then triumphantly said, “See?”
What I saw after that brief pause was a page of relevant cases, statutes, regulations, blog posts, judge-crafted summaries, briefs, and more. I learned that CARA’s artificial intelligence used something called “latent semantic analysis” to pull context from my document and build connections to my issue. The tool found patterns, not simply words. And it did it faster and better than I could.
With a drag, a drop, and a general concept, I had what I needed to begin crafting arguments.
I immediately saw the significance of this technology. To my surprise, I was mad. I’d spent years in law libraries and my tiny office, stressing over which magical search term might result in the case I needed. It suddenly seemed like such a waste.
CARA found the pattern in seconds on my home computer screen, which stared blankly back at me, asking what I’d do with it. I stared back with surprise and anxiety. I used to spend all my time navigating words in search of a barely-informed pattern. This thing did it better in seconds. What now?
Maybe that’s how the creators of Stranger Things felt. I imagine a meeting of artists suddenly interrupted by Netflix’s data scientists pouring out a list of must-haves. Didn’t the machines just do all the work?
If the technology can find patterns that well and that quickly, what’s left for artists?
What’s left for lawyers?
Legal Research in the Age of AI
Mackenzie Dunham runs a “low bono” firm in Texas called Access Justice Houston. He agreed to hop on a call with me to explain how he uses CARA as an ally rather than a replacement for his skills.
Mackenzie told me about a sexual assault case that had really challenged him. Petitions for sexual assault protective orders are unfortunately common for family lawyers like Mackenzie, but they’re based in multiple sources of law. He had a big haystack in front of him and needed a very particular needle.
Mackenzie had to find an authoritative definition of “reasonable grounds to believe.” After struggling with secondary sources and traditional search methods, Mackenzie looked to Casetext’s advanced legal research system. He wasn’t optimistic.
“My co-counsel and her entire organization had been looking for the same thing for like a year and a half,” Mackenzie told me.
Using Casetext, however, Mackenzie did find something: “Literally, the first thing that popped up in Casetext [was] a U.S. Supreme Court case analyzing what a ‘reasonable ground’ meant and comparing it to probable cause.”
Casetext’s artificial intelligence found Mackenzie’s crucial result because it didn’t suffer from his limitations. CARA looked outside the family law context while he and his colleagues had limited their searches to that area. The AI, armed with the context of his document and questions provided, scoured its entire library and found Mackenzie’s definition in criminal law. Only a machine could sort through such a large haystack that quickly.
With the data in hand, Mackenzie got busy discovering the insight that applied to his case. He crafted an argument that won over the judge.
Casetext’s Vice President of Business Development, Anand Upadhye, tells a similar story. During a demonstration to a firm, the lawyers asked if they could drop one of their most troublesome cases into CARA. When the first result for their corporate breach of fiduciary duty search was a divorce case, Anand tried to quickly scroll past it.
Fortunately, the lawyers insisted that he open the case and dig in. Why would a divorce case be relevant? The training they’d received on navigating library bookshelves couldn’t provide an answer.
As it turns out, the divorce case involved a couple who owned a corporation. The judge was asked to rule on whether their split created a breach of fiduciary duty. Much of the judge’s reasoning overlapped with the lawyers’ pressing questions. They immediately changed their approach, edited their brief, and added the divorce case. It proved vital.
Without preconceptions or distractions, and with the ability to pore through enormous data sets, the technology did what these lawyers couldn’t. It found the pattern.
Faced with that kind of speed, efficiency, and accuracy, Mackenzie and the corporate lawyers might have feared replacement. Instead, they signed up for Casetext accounts.
Why? What was left for them to do?
What AI Leaves for Lawyers
To put these lawyers’ process in perspective, think back to Stranger Things. The algorithms did not create that show. They couldn’t have.
In fact, the very human creators of Stranger Things (twin brothers Matt and Ross Duffer) obsessively pursued the project. They received 15 rejections from studios before Netflix jumped on board. They simply needed the algorithms to validate and improve their creative effort.
The Duffers asked a simple question: Can we make a nostalgic monster series with child protagonists? and the algorithms answered yes. The data revealed a pattern that served the artists and the company.
The AI then directed the Duffers on how to turn their idea into a blockbuster. They worked in tandem with the data tools, not against them.
Other studio executives had missed out on the hit show because they couldn’t see past their own biases. They told the Duffers that they either needed to make the show less scary or remove the kids. The executives failed to see the pattern.
This contrast reveals a successful process that you can implement:
- Use creativity to ask the right questions,
- Put your questions to the most helpful tools, and
- Incorporate the patterns those tools discover to guide your creative work.
Your humanity comes in both before and after the data tools do their work. That’s where you’ll create the value that matters to clients.
Despite the mythology around technology replacing us, clients care about insights. They need the truths only humans can synthesize and explain. But to get to those insights, we first must find patterns. The bucket of tools we call “artificial intelligence” can help us with that. We just need to use them.
Enabling the Good Stuff
Advanced technology doesn’t eliminate the need for creativity. It simply defines constraints that will channel and optimize our creative efforts. For now, these tools need us as much as we need them.
Still, we must balance the need for a human touch with openness to better ways of operating. Resist the temptation to behave like Netflix’s competitors. They missed out on a blockbuster because they guessed. They limited their data set to their own biases and experiences. In 2019, we have better options than our memories and magic keywords.
AI tools allow creators like you to ask questions, discover patterns, and focus on crafting insights. The technology not only leaves us that crucial work, but it also can’t do it without us.
So what’s left for lawyers who use AI as an ally?
Give these tools a try. Then go and make your art.
About the Author
Mike Whelan, Jr. is Managing Editor at Casetext. He writes free, comprehensive practice guides for solo lawyers at casetext.com/blog.