AI, Technology and the Law

AI, Technology and the Law Conference: Charting the Emerging Debates in Legal Technology

By Ellie Marshall

On Saturday, March 25th the Centre for Innovation Law and Policy hosted an impassioned and thought-provoking conference on the interfaces between legal technology and legal theory. Sponsored by the University of Toronto Law Journal and the Social Sciences and Humanities Research Council, and organized by Professor Simon Stern, the event covered topics ranging from how artificial intelligence will alter the practice of the law, to how technology can democratize the justice system and legal education. 

The conference began with a panel entitled “How Will Artificial Intelligence Alter the Practice of Law,” moderated by Michael Hilliard, of Inter Alia Law. To kick-off the day’s discussion, Hilliard recalled his own early-career experiences in the legal industry, where he was witness to what he thought was “game-changing technology” before suggesting that AI is the first technology that could actually truly implicate the delivery of legal expertise. Panelists Ben Alarie and Albert Yoon picked up this theme in their presentations, which focused on insights gained from their own legal technology startup, Blue J Legal. Alarie, asked the audience to think about artificial intelligence not as a scary “Terminator style future” but rather simply as applied mathematics. “We will have more transparent and predictable law because there will be more certainty,” claimed Alarie, referring to tools like Blue J, which are able to parse entire corpuses of legal decisions to point users to the most likely judicial outcome for a specific set of facts. Yoon, building off this idea, suggested that "AI will reduce costs to the client, shifting demand curve for legal services down, meaning more people can get legal services." Hilliard challenged these claims, and their impact on the legal industry, by suggesting to the panelists that to be a good legal strategist, lawyers will still need to do their own  legal research. Alarie’s response indicated that, in his view, the ideal use of machine learning based legal tools would be in concert with a “human lens.” Similarly, Yoon suggested that machine learning might not be appropriate in scenarios where the law is less developed, such as animal rights.

The second panel, “Will Technology Democratize Access to Legal Services,” chaired by distinguished Prof. Martin Friedland, began with Sherry MacLennan of British Columbia’s Legal Services Society. MacLennan provided an overview of important legal technology developments which seek to streamline the judicial process, usually for self-represented litigants. MacLennan highlighted her organization’s MyLawBC (http://www.mylawbc.com), which provides “legal pathways” to citizens facing family and estate planning disputes. Frank Pasquale’s presentation continued the first panel’s discussion of potential built-in bias in computer systems by suggesting that too much credit may be given to the programmers of legal technology applications and not to the citizens to whom the majority of the labour is shifted. Pasquale reminded that audience that technology has the potential to further access to injustice. Pasquale further cautioned academics to not overemphasize the disruptive potential of legal technology and focus discussion on the emerging issues of algorithmic accountability and privacy concerns. Paul Gowder complemented Pasquale’s point by highlighting that “lawyers supply class-based credibility & other non-legal cognition based advantages to subordinated.” Gowder emphasized that legal technology risks eroding support for the underclassed by creating tiers of support, where those with resources would continue to access traditional legal resources. The conference attendees quickly grabbed onto this point and asked whether legal technology does not simply shift liability around.

 
Dana Remus, keynote speaker

Dana Remus’ keynote presentation (https://www.youtube.com/watch?v=UYSZeHqZnaA) continued on the theme of questioning technologically determinist approaches to the law. Remus began by explaining how machine learning tools like IBM’s Watson, may struggle with the law because “to automate a lawyering task it needs to be possible to articulate it as a set of rules or instructions.” Further, “at the present time unstructured human interaction permeates every level of the legal industry, resistant to automation.” This direct response to Prof. Alarie’s positivist claim that the law can and should be fully determined, emphasized on the importance of meaning in the law. “Data processing is still based on linguistic statistics and not actual meaning.” Remus also noted that “access to justice is a key obligation of the legal profession and it cannot be solved by redefining it as access to technology.” Remus concluded the keynote by encouraging legal professors to engage their students in the process of building legal technology tools so that students are able to see what aspects of the profession can and cannot be automated.

The third panel comprising of Michele Pistone and David Thomson entitled “How Will Technology Change Legal Education” discussed the impact of technology on legal education. Pistone noted that technology opens doors for new ways of conceptualizing & designing education. She suggests that after we identify pockets of non-consumption in legal education, law schools must determine how they can effectively respond to fringe student communities. Using the example of her own work at Villanova Law School to democratize education around refugee law, Pistone advocated for the benefits of measuring student competencies instead of statistical indicators such as test scores and GPA. Thomson echoed Pistone’s concerns and pointed to three intractable problems in legal education: the mismatch between available jobs and need for representation, skyrocketing tuition & embarrassingly low levels of diversity.

The final panel of the day, “Will Technology Challenge the Conceptual Foundation of the Law” featured Mireille Hildebrandt, Brian Sheppard and Daniel Katz. The panelists focused on how we can use existing legal theories to conceptualize technological changes in the law. Hildebrandt posited four concerns with the growth of artificial intelligence: the incontestability (or “black-box nature”) of predictions, a problematic shift from decisions based on reason to decisions based on statistics, a potential “de-skilling” of lawyer, and the potential for technological applications to infringe on fundamental rights such as privacy and the presumption of innocence. Sheppard added to this by asking whether the features make AI attractive are in contest with the concept of human-ness in our legal system. Katz offered a rebuttal to these two positions by emphasizing the market forces which are pushing the use of artificial intelligence, regardless of legal foundations. 

Conference papers will appear in a special issue of the University of Toronto Law Journal to continue this exciting conversation. The journal is scheduled for release next year.