The most widely read magazine for Canadian lawyers
Issue link: https://digital.canadianlawyermag.com/i/1499770
FEATURE 12 www.canadianlawyermag.com CROSS EXAMINED internally from their own teams, they're looking for that from external legal counsel." The second big challenge, Veel says, was ensuring they were inputting good-quality data. "We've got clients who expect good predictions if we're going to give them these numbers. And so, we're similarly obsessed with data quality, which means that it takes a long time to build up the dataset." So, Veel and his team of lawyers, articling students, and undergraduate interns spent a lot of time looking at cases, identifying infor- mation about those cases, and plugging them into a dataset. "One of the little secrets of a lot of data analytics programs is that the vast majority of the work is spent on building your data sets on the front end. You can get very robust predic- tive models in a matter of seconds from your data," he says. Computer coding took up much less time for the Lenczner Slaght team. Veel did not study computer science but did much of the coding himself. "You end up out of necessity doing some coding in various statistical software packages." Despite its unproven approach, the firm encouraged Veel's investment in the project. "One of the things I love about Lenczner Slaght and [one of ] the reasons that I have been here for the last 13 years is that there is a culture of innovation, a culture of exper- "We've got clients who expect good predictions if we're going to give them these numbers. And so, we're similarly obsessed with data quality, which means that it takes a long time to build up the dataset" imentation. [There is] a lot of support for new ideas and new initiatives and a recog- nition that any good project is going to need some initial outlays and that it's okay if some percentage of those projects fail." In March 2021, the firm officially launched its data-driven decisions program, encom- passing data from several courts and tribunals. "We are getting to the point where those types of models can be good enough to be useful complements to the reasoning that lawyers employ," says Veel. With the recent renewed interest in arti- ficial intelligence and its uses in the legal profession, Veel's work has taken on new urgency. Still, he is quick to differentiate it from tools such as ChatGPT. "They're not language models. They're more in the structured datasets that academics historically used. What we're doing is not fundamentally dissimilar from what empirically minded social science researchers have been doing for decades." Veel stresses that this initiative is less about speeding up his team's work and more about finding new insights that can challenge the profession's conventional wisdom. "It doesn't take the lawyer out of the equa- tion, but it provides an additional data point to either confirm or interrogate what we know. It's effectively like another opinion you're getting from an unbiased external observer that can help us as the lawyer give better advice." DATA-DRIVEN DECISIONS Lenczner Slaght is currently collecting and analyzing empirical data for the following projects: Supreme Court of Canada decisions Supreme Court of Canada leave decisions Ontario Court of Appeal decisions Federal Court of Appeal patent cases Class actions Commercial List of the Ontario Superior Court Competition Tribunal cases