Choosing an AI tool a 'know thyself' moment for legal professionals

This article was produced in partnership with Thomson Reuters

Lawyers know what they’re doing: they’ve got the training, they’ve got the experience, and over time, they’ve developed their own ways of working. New technology, of any sort, is a disruptor to the unique processes of a group not known for its immediate willingness to innovate. But the onslaught of artificial intelligence (AI), arguably the most disruptive technology to ever hit the market, has been overwhelmingly well-received — and that’s a big clue when we look ahead, says Thomson Reuters’ Ryan Groff.

“This technology hit like a bang and legal professionals have jumped at it, drawn like moths to a flame,” Groff, senior solutions consultant, says. “That’s an interesting indication of what this technology might mean for the future of law.”

But while AI’s phenomenal capabilities have behooved many in the legal profession to adopt the powerful new tools as quickly as possible, they shouldn’t forget to think clearly about why they’re implementing it, he adds.

“They must fundamentally examine their workflow, which may not even exist in a stated way but more as professional muscle memory. This is a real ‘know thyself’ moment that forces everyone — whether it’s a firm, a sole practitioner, or the in-house legal department of a multi-international company — to take the same step and ask: ‘What is the exact thing I’m trying to solve for?’”

Setting the stage: AI as assistant

This isn’t the first time a profession has encountered transformation, but there’s never been a technology that understands human language the way today’s AI does. For a discipline that focuses on words all day long, every day, it surfaces “some really interesting possibilities for the practice of law,” Groff says.

Specifically, lawyers can move away from the mechanics of a keyword where they throw one out, evaluate the results, and keep hitting their head against the wall until they find what’s relevant, to a tool that’s able to articulate entire concepts. AI can be told to look for an idea or an abstract idea such as an emotion, like “unease,” without the user sitting for hours guessing the keywords that represent this feeling. 

As AI’s capabilities grow in leaps and bounds, at this stage it’s the pace of change that’s distinctive — and can often feel disorienting.[1] Groff teaches a course on law innovation and technology and sees firsthand how this affects even the newest lawyers.

“My students frequently talk about the uncomfortableness of being on a staff with people who have many different technical aptitudes and the stress they feel about keeping up with the changes,” he notes. “Remaining current on new technology is a lot. There’s a layer of patience and understanding we need to have for ourselves and each other.”

For all of AI’s remarkable abilities, it’s critical to set the stage: it’s an assistant, not a replacement. The disorientation that comes with the pace of AI advancement can make it easy to jump to the conclusion that it’s taking people’s jobs, but that requires a closer look, Groff notes.

It’s true that AI users have the technology to automate lower-level tasks, but “most of that replacement conversation is a hype train that needs to be thought through very carefully.”

In his role, Groff helps people understand what Thomson Reuters’ CoCounsel is capable of as well as its limitations, so they can build value for themselves through informed decisions over what technology to adopt. It’s common that people ask how they can ensure the tool is always right, and Groff explains that it’s all in the name: an AI assistant for legal professionals.

“I remind them we didn’t create a god; it doesn’t read minds. Instead, we offer a tool that can read, write responses, and follow instructions better and faster than prior technologies. When it comes to research, tools like CoCounsel deliver answers to entire ideas in a way that prior research technology could not.” “What is it, an assistant or an AI lawyer? It’s clearly not the latter — it’s assisting you.”

Groff also tells the lawyers he meets with that the very skills that make them extremely effective today, are the ones needed to oversee the assistance they receive from AI or a more junior human colleague, for that matter. They should ask enough questions to feel satisfied with the work in front of them, standing as a backstop for AI and not the other way around. Lawyers have always used assistance in this way, and nothing about today’s AI changes this dynamic.

Those techniques are even more important when working on AI’s curve of development, because lawyers need to much more quickly know how to use their professional experience to assess the tool’s output. Faster answers require quicker, keener analysis from overseeing lawyers.

“In many ways, it actually draws us in a much more focused way into our training, because we have to use it at a pace that we may not be used to,” Groff says.

That said, it’s equally important to understand where the tool you choose is drawing its information from. A lot rests on the quality of the data feeding the tool. A highly curated database of legal content — such as the one Thomson Reuters has been carefully building for the better part of 150 years — is going to be far more reliable than if the AI is scanning the open internet.

In the same vein, a team of lawyers builds the interface between powerful Large Language Models (LLMs) and the legal professional end-users. As a result, when CoCounsel is asked to help,  whether by a lawyer in Canada or Kansas, the result will be of a calibre that a lawyer would expect.

“That sounds easy in one sentence, but that’s the very important work,” Groff stresses. “And that’s the special part of CoCounsel: a team of lawyers, AI product developers, and machine learning scientists with years of experience marshaling all the power of these LLMs to be used within the legal sphere. You need that kind of expertise behind the AI you choose.”

AI continues making its mark on the legal industry

Groff was part of legal startup Casetext before it was acquired by Thomson Reuters last year and is overjoyed to see AI “become what we thought it could be.” But there’s no summit in sight, he adds. It’s a continuous evolution that brings about new capabilities to grapple with, and because legal professionals have an ethical requirement to understand any significant new technology that hits the market, “it can become almost irresponsible to not at least dabble,” Groff says. Part of this transformation requires “a playfulness” and willingness to step outside the comfort zone of established processes.

Groff predicts AI will soon be able to “think with you” and connect the dots from task to task on its own, rather than having to be asked to complete each step individually. The user will tell the tool their end goal and the technology will use its own skills or the other apps it’s connected with to accomplish an entire workflow.

AI’s potential is almost limitless and if you’re still searching for the right tool — and Thomson Reuters brings to the table a top contender in CoCounsel, Groff adds — it’s time to do the work to understand what you require. And Groff and his colleagues are here to help separate the wheat from the chaff.

“At Thomson Reuters, we’re building on over a century of expertise in the legal market, from paper-based books of caselaw, to online resources, and now an AI tool that’s tailor-made for the profession. We’re happy to help you determine what you’re trying to solve for with this technology.”