AI musings and redactions run wild

A summary of two varying views on the potential of AI in the tribunal/court setting and a case of extreme redactions

Winter has arrived in the nation’s capital this week - about 10 cm of snow. I broadcast a show featuring the experimental composer, Alvin Curran, including an interview with the soon-to-be 86 year old composer from his home in Rome.

Part of my webinar on building inclusion in the hearing room has been transformed into a micro-learning video on pronunciation of names. Stay tuned for more mini videos.

Nominations time for the 19th annual Clawbies

These annual awards for “the best, most original, and most influential Canadian legal content on the Web” have been in place for 19 years. That includes blogs, newsletters, podcasts, videos and social media accounts.

This year, nominations will be through BlueSky and LinkedIn.

Here are the “rules”:

Nominate up to three digital publications or authors (remember to use the #clawbies2024 hashtag) that are freely available, at no monetary cost. Include a brief explanation of why you think those publications deserve an award.

Nominations will be accepted until the end of day on Monday, December 16th, 2024. Then stay tuned for the big reveal on New Year’s Eve.

AI and tribunals - optimistic and pessimistic views

Paul Daly has written a mostly optimistic view of the role of AI in the tribunal setting:

…The very thing that artificial intelligence promises – efficiency and effectiveness – is exactly that which administrative tribunals have long primarily been designed to achieve. This is not to suggest, however, that administrative tribunals should move blindly forward and embrace AI in all shapes and forms. In Part I, I also suggest that careful, contextual analysis is required before using AI and lay out several typologies that are helpful in conducting the necessary analysis.

In Part II, I then discuss the ways in which administrative tribunals can use AI: communicating with the public and enhancing the accessibility of their websites; assigning decisions more effectively to members of the tribunal; triaging files to ensure decisions are made as quickly as possible; drafting decisions; and, conceivably, making decisions. These use cases are set out in increasing order of audacity. Machines making websites more usable for individuals is one thing, making life-defining decisions about individual rights and obligations quite another. In the end, though, I will argue that there is a broad range of potential uses for AI in administrative tribunals.

His University of Ottawa colleague, Amy Salazyn has a decidedly more pessimistic view:

My focus in this column will be on some of the more subtle ways that generative AI risks undermining confidence in our justice system and introducing error and bias into our court processes. I characterize these as generative AI’s potential corrosive risks. The possible harms arising from the risks I canvass below aren’t as immediately dramatic or legible as the harms resulting from AI generating fake cases/evidence or from introducing AI systems that take over the judicial decision-making function. The claim here is, however, that these corrosive risks have the potential to create a “rust” which may, particularly as it collects and accumulates, have some very negative effects on our justice system.

…

But, in order for our courts to fulfill their important role in our democracy, they need to have healthy fact-finding and law-making/law-applying processes. If courts start misstating relevant facts or the law, even in subtle ways, or if the work of the courts starts becoming overly influenced by choices of private legal tech developers, the administration of justice and the public’s confidence in courts risks being undermined.

So, what to do? The answer isn’t, in my mind, for courts to avoid exploring how generative AI technology might improve the administration of justice. This technology is still new and we don’t yet have a clear and full picture of how we might productively and ethically engage with it. What we need to do, however, is move forward cautiously and with our eyes fully open.

I’ll end with three very quick suggestions. First, building awareness about some of the more subtle risks of generative AI discussed above should be a priority. Awareness is particularly important given the low barriers to access for many AI tools and the prospect for “shadow AI” when it comes to assistive uses (i.e. the potential that some individual judges may use certain AI tools even if they have not been formally vetted or approved).

Second, the concerns raised above suggest that significant caution is warranted when it comes to court use of generative AI even with use cases that do not directly intersect with judicial decision-making. There can be a tendency to view “assistive” or “internal” uses of generative AI by the courts as low-risk and therefore not warranting deep scrutiny. Courts ignore the risks of such use cases at their peril.

Finally, the discussion here links to a broader conversation about the need for more transparency and better benchmarks when it comes to generative AI tools and issues of bias and accuracy. We need a clearer understanding of how these tools work, what their limitations are, and what sorts of preferences or priorities are built in.

Redactions run wild

This recent decision from the Federal Court on informer privilege is a bit difficult to follow.

Cases of note

Tribunal must explicitly consider medical report contrary to its ruling

In Dmitrienko v. Canada (Attorney General), 2024 FCA 194, the applicant had been denied compensation for a medical condition by the Veterans’ Review and Appeal Board (VRAB) because the panel found no connection between his compensable condition and the claimed disability. Most of the medical reports found no connection, but one report found that there was. The panel did not refer to this outlier in its decision. The Federal Court of Appeal held:

[7] While administrative decision makers may, in some circumstances, be presumed to have considered all of the evidence before them, that presumption cannot operate to save the Appeal Panel’s decision in this case. This is because the Supreme Court of Canada has told us in Canada (Minister of Citizenship and Immigration) v. Vavilov, 2019 SCC 65 that administrative decision makers must provide reasoning on key issues affecting the outcome of a case. In addition, section 39 of the Veterans Review and Appeal Board Act, S.C. 1995, c. 18, instructs the VRAB to draw every reasonable inference in favour of an applicant based on the evidence presented to it, and to resolve any doubt in favour of the applicant.

[8] The Appeal Panel had evidence before it that supported Mr. Dmitrienko’s claim that his ongoing cardiorespiratory symptoms resulted from the PE that he suffered following his knee surgery for his on-duty injury. While the Appeal Panel was not required to accept Dr. Stuart’s opinion as to the linkage between these events (and indeed, the Review Panel appears to have had concerns with respect to this report), it had to at least come to grips with it in determining whether Mr. Dmitrienko had established a case. Having failed to do so, the Board cannot be said to have fulfilled its obligations under Vavilov and section 39 of the Act.

…

[10] Administrative decision makers do not have to deal with every issue raised by a case, however subordinate, collateral or incidental. However, the causal link between Mr. Dmitrienko’s PE and his cardiorespiratory symptoms was the central issue in this case. The failure of the Appeal Panel to engage with Dr. Stuart’s evidence means that the reconsideration decision lacks the lacks the transparency, intelligibility and justification required of a reasonable decision: Vavilov, above at paras. 96-98, 127–128.

[11] The Federal Court addressed Dr. Stuart’s report at paragraphs 68-71 of its decision, explaining why, in its view, the report was entitled to little weight and would not have made a difference to the Appeal Board’s decision. It is not, however, the role of the Federal Court to weigh the evidence and make findings of fact to buttress an administrative decision. That responsibility rests with the administrative decision maker, and “even if the outcome of the decision could be reasonable under different circumstances, it is not open to a reviewing court to disregard the flawed basis for a decision and substitute its own justification for the outcome”: Vavilov, above at para. 96.

Expeditious hearings not procedurally unfair in high-volume tribunal

In Casiechitty v. Imran, 2024 ONSC 6751, the landlord alleged that the hearing before the Landlord Tenant Board was procedurally unfair. The court found the allegation to be without merit:

[7] … Requirements of procedural fairness must be determined in the particular context in which they arise. Tribunals have authority to control their own process and courts will accord deference to those choices, as long as they are not procedurally unfair.

[8] The Board is a high-volume expert tribunal. It is required by statute to adopt the most expeditious method of determining questions arising in a proceeding that affords all parties directly affected by the proceeding and adequate opportunity to know the issues and be heard on the matter: Act, s. 183. The Board’s rules permit the Board member to define and narrow the issues, question a party or witness, and limit the evidence or submissions on an issue.

[9] In this case, the landlord’s two applications were heard over two days and the tenant’s interrelated applications were heard over a day. By the third day, it was reasonable for the Board member to insist on moving the matter along given the resources that had already been devoted to the matter and because it had not proceeded efficiently. The Board found the landlord’s representative to have caused extensive delay and disruption to the point that costs were ordered against the landlord.

Links of interest

A day at Quebec's jammed rental tribunal

The strain on resources is visible at TAL, with a steady stream of people arriving at the offices throughout the day. A security guard is ready to intervene when frustrated tenants raise their voice.

Lemieux said more resources — and better prepared landlords and tenants — would make the process work better.

"If ever a case starts to sound like it's going to take longer, the judge is like, 'we've got to postpone it and hear you in front of another judge," Lemieux said.

"If the TAL were to be able to, you know, have more judges so that way people can have more time and then not have to wait two months to be able to get a hearing, that would be beneficial."

The Application of the Charter to Ontario School Boards

In York Region District School Board v Elementary Teachers’ Federation of Ontario, 2024 SCC 22 [YRDSB], the Supreme Court of Canada (“SCC”) reviewed an arbitrator’s decision regarding the reasonable expectation of privacy of two teachers at an Ontario public school. The majority conducted a correctness review to set aside the arbitrator’s decision, holding that the Charterapplies to Ontario public school boards and that the teachers’ privacy rights were breached.

…

The SCC issued a signal defeat for the possibility of an important Charter protection of students in Canadian Foundation for Children, Youth and the Law v Canada, [2004] 1 SCR 76, in which it upheld the constitutionality of s. 43 of the Criminal Code, RSC 1985, c C-46. That provision permits a teacher’s “reasonable” use of corrective physical force. YRDSB may perhaps unintentionally subject teachers’ other disciplinary actions, authorised under a different legal framework, to additional challenges.

Constitutional Limits on the Administrative State

This is a very useful, well-written decision that touches concisely and incisively on some key basics of administrative and constitutional law about the limits on the authority (or, dare I say it, “jurisdiction”) of administrative decision-makers.

Commercial arbitrators and conflict of interest

The Aroma appeal clarifies the standards applicable to determining whether justifiable doubts over impartiality or a reasonable apprehension of bias by an arbitrator exist. This clarification is important for counsel and arbitrators, particularly as law firms frequently engage the same arbitrators for different disputes.

Aroma explains that if a party or their counsel has particular concerns about an arbitrator taking on other mandates during the course of the arbitration, those concerns should be identified and clearly disclosed to the arbitrator and opposing counsel prior to the potential engagement. Alternatively or in addition, the parties can incorporate a more stringent standard for disclosure of potential conflicts, such as the subjective test set out in the IBA Guidelines.