Generative AI is a powerhouse of potential, providing creative ways to add value to the services lawyers provide. But despite much conjecture over what it may be capable of, as it stands it shouldn’t serve in any way as a shortcut – especially with the high-end legal work involved in complex commercial litigation, warns Alexandre Baril-Furino, partner at Woods LLP.
“Technology advances quickly but gen AI isn’t at a level to do the lion’s share of the lawyer’s job,” Furino notes. “It’s more about using it to think outside the box once you’ve done the work the good old-fashioned way.”
Furino’s advice when working with AI tools? Be wary.
“Ask for sources, verify everything, take nothing for granted.”
Acting as a backstop
While AI is not novel in the litigation space – it’s served lawyers well in legal research and document management for some time now – usage is evolving. Instead of key words, lawyers can describe what they’re looking for, and AI sources caselaw or authorities that fit that context. AI is effective at organizing documents, identifying responses in discovery, or generating lists, and can also generate a range of documents, whether it be memos or analysis, outlines of arguments, examination outlines, or timelines.
But there are limits to what gen AI can produce, and its ability to create content raises the stakes. No matter how it’s leveraged, lawyers still must act as a backstop by reviewing every output, guarding against hallucinations or other inaccuracies.
“Lawyers’ work is still fundamental in producing high-quality documents,” Furino says, adding that while some clients associate AI with reduced costs, it’s not always about saving money but achieving the best results.
In document management, AI can alleviate the need for a team of lawyers and paralegals to pore over the information, saving time and client dollars, but using it to create memos or other written submissions doesn’t have the same return. Though output quality varies depending on the precision of instructions provided – and legal professionals continue to sharpen prompting skills – as it stands, anything AI generates serves at best as a first draft. But Furino believes it’s most impactful to do the work as usual and apply the technology afterwards to refine it, while prioritizing data security standards above all else.
“It could provide an analysis with the same data for comparison, be prompted to challenge the position or solution you came to, surface an angle you didn’t consider, or highlight information to use more extensively,” he explains. “I caution very much against thinking AI serves as a shortcut to the end. It brings more value to the client when, on top of your expertise, you use it to verify or counter-verify.”
Identifying trends
As gen AI use becomes more prevalent in the litigation space, there are some notable trends on the rise; for example, applying predictive analytics to litigation outcomes, which Furino calls “ambitious.” There are many unpredictable elements that affect the outcome of a litigation, and, given that AI tools work based on the information provided, it’s not realistic to expect an accurate predictive analysis – at least for now.
“Every lawyer knows you may have a good case on paper, but it can all change in the courtroom,” he says, adding that his stance extends to the use of AI by judges to aid decision-making. He points to other countries that have regulated its use in the courtroom and are leaning away from this sort of automation.
“Even if it could be done, it doesn’t mean it should be done,” Furino notes. “There’s an element of discretion that falls within the decision-maker’s purview that can’t be done by AI. Parties also take comfort in having an experienced, qualified, and honourable judge listening to and deciding their case.”
Litigation related to AI is to be expected, especially in Canada where there’s no specific regulatory framework. Grappling with considerations related to privacy, consumer protection, and IP rights, among others, the emerging area is rife with “loose ends and blind spots.” When parties are contracting, they’re not able to include clauses to deal with situations they haven’t foreseen, creating an environment more conducive to litigation or “a bit of a Wild West.”
While the landscape continues to settle and expand, Furino sees gen AI evolving rapidly with respect to the reliability of output and lawyers learning to use the tool better. But he doesn’t foresee a time where it will surpass human lawyers, especially when it comes to complex commercial litigation.
“We look at problems in front of us with skill, competency, and training and in a way that keeps in mind clients’ objectives and many other considerations. I’m not convinced that can ever be replicated.”