When AI Starts Making Administrative Decisions
Marco P. Falco
Is Virtual Justice Real Justice?
When AI Starts Making Administrative Decisions
Marco P. Falco
Originally published on Law360 Canada, formerly, The Lawyer’s Daily, © LexisNexis Canada Inc.
One of the tenets of procedural fairness is the right to reasons for an administrative decision. Canadian tribunals and decision-makers have an obligation to explain why a particular outcome is reached and the rationale for it.
The use of artificial intelligence (AI) as a tool in tribunal adjudication is challenging the common law standard for what amounts to adequate reasons for decision.
For example, does the right to reasons established in the Supreme Court of Canada’s 1999 decision, Baker v. Canada (Minister of Citizenship and Immigration) ,  2 S.C.R. 817,  S.C.J. No. 39 (S.C.C.), include the duty that a human tribunal member draft or arrive at the decision in question? To what extent can a tribunal delegate its decision-making authority to AI in the interests of efficiency and to avoid delay? Will courts scrutinize the inherent or potential biases of AI that could taint a particular tribunal result?
Most of these questions remain unresolved.
A recent decision of the Federal Court, Haghshenas v. Canada (Minister of Citizenship and Immigration) ,  F.C.J. No. 432, 2023 FC 464 (F.C.C.), sheds light on how Canadian courts may approach the fairness or reasonableness of administrative decisions written with the assistance of AI.
A POWERFUL CHINOOK
Haghshenas involved an application for judicial review of a decision by an immigration officer (the officer) at the Canadian Embassy in Turkey. The officer denied the applicant a work permit designed for entrepreneurs and self-employed foreign nationals seeking to operate businesses in Canada (the work permit).
One of the requirements for the work permit under paragraph 200(1)(b) of Canada’s Immigration and Refugee Protection Regulations, SOR/2002-227 (the Regulations), is that the officer be satisfied that the applicant “will leave Canada by the end of the period authorized for their stay.”
In this case, the officer concluded that the applicant would not leave Canada at the end of his stay under the work permit. That is, the applicant’s intended aspiration of starting an elevator/escalator business in Canada did “not appear reasonable” given the speculative revenue projections for the business and the fact that the company had not obtained the appropriate licences, among other reasons.
In reaching this decision, the officer employed Chinook, a Microsoft Excel-based tool developed by Immigration, Refugees and Citizenship Canada (IRCC).
According to the IRCC website, Chinook helps with “temporary resident application processing to increase efficiency and to improve client service,” with the goal of assisting in the backlog of work permit applications. It “does not utilize artificial intelligence (AI), nor advanced analytics for decision-making, and there are no built-in decision-making algorithms.”
These statements notwithstanding, the applicant challenged the officer’s use of Chinook on judicial review, arguing that employing AI to reach an administrative decision was both procedurally unfair and substantively unreasonable.
The Federal Court dismissed the applicant’s position, largely on what appears to be an assumption that the Chinook tool constitutes a form of AI. In so doing, the court hinted at a number of important principles about how it may scrutinize the use of AI in administrative decision-making in future.
1. DECISIONS MADE BY HUMAN DECISION-MAKERS ARE NOT PROCEDURALLY UNFAIR
In rejecting the argument that the use of AI was procedurally unfair, the court appears to have drawn a line in the sand about the proper role of AI mechanisms in administrative decision-making. The court noted that in the applicant’s case, AI did not reach the final decision regarding his work permit — the officer did.
Inherent in the court’s reasoning is the presumption that it is procedurally fair for AI to assist an administrative decision-maker in rendering reasons for decision.
Inherent in the court’s reasoning is the presumption that it is procedurally fair for AI to assist an administrative decision-maker in rendering reasons for decision. AI assists the administrative state in the goal of promoting more efficient and timely outcomes.
What appears unfair, however, is the state’s delegation of its decision-making authority to AI. The court held:
As to artificial intelligence, the Applicant submits the Decision is based on artificial intelligence generated by Microsoft in the form of “Chinook” software. However, the evidence is that the Decision was made by a Visa Officer and not by software. I agree the Decision had input assembled by artificial intelligence, but it seems to me the Court on judicial review is to look at the record and the Decision and determine its reasonableness in accordance with [the Supreme Court of Canada’s decision in] Vavilov . Whether a decision is reasonable or unreasonable will determine if it is upheld or set aside, whether or not artificial intelligence was used. To hold otherwise would elevate process over substance.
2. IT IS NOT UNREASONABLE TO USE AI IN ADMINISTRATIVE DECISION-MAKING
Separate from the issue of whether the use of AI is procedurally unfair, the court also rejected the argument that the officer’s reliance on Chinook rendered the decision substantively unreasonable.
According to the court, there is nothing inherently unreliable or ineffective about the use of AI, at least in this particular case. The court did not deem it necessary to delve into the inner workings of the Chinook software to determine if its mechanics were inappropriate or would lead to unreasonable results in the immigration assessment process:
Regarding the use of the “Chinook” software, the Applicant suggests that there are questions about its reliability and efficacy …. [T]he Applicant suggests that a decision rendered using Chinook cannot be termed reasonable until it is elaborated to all stakeholders how machine learning has replaced human input and how it affects application outcomes. I have already dealt with this argument under procedural fairness and found the use of [AI] is irrelevant ….
So, in this particular context, the government’s use of AI survived reasonableness scrutiny.
WILL AI REPLACE TRIBUNAL DECISION-MAKING?
The court’s approach above reflects a willingness to accept machine learning as a limited component of administrative decision-making, with the caveat that ultimate adjudicative authority must reside in a human tribunal.
Haghshenas just scratches the surface, however, of the implications of the Canadian administrative state delegating its roles and responsibilities to machine learning in the interests of efficiency.
As we are learning, AI comes with its own set of inherent biases and problems.
There will no doubt be new circumstances in which an enterprising lawyer will argue that a tribunal’s reliance on AI tainted the outcome of the decision or rendered it procedurally unfair. Tribunals and agencies across Canada must, therefore, approach the question of whether and how to adopt AI in the decision-making process with a degree of caution and with significant legal and ethical training. This is the only way to ensure that the use of AI remains a fair and reasonable tool in administrative adjudication.
Marco P. Falco is a partner in the litigation department at Torkin Manes LLP in Toronto, who focuses on judicial review and civil appeals. He is also co-chair of Torkin Manes’ diversity and inclusion committee.
Politicking Around Prisoner Classification Displaces Reason with Rhetoric
Originally published on Law360 Canada, formerly, The Lawyer’s Daily, © LexisNexis Canada Inc.
Former Public Safety Minister Marco Mendicino recently issued a public statement on the transfer of Paul Bernardo to a medium security prison. This was done in response to public confusion and outrage about why the transfer had taken place. But instead of explaining the classification framework used in federal prisons, the former minister instead remarked that the decision was “shocking and incomprehensible,” and vowed to intervene.
Bernardo, as someone who in sentencing parlance would be considered “the worst offender in the worst circumstances,” is understandably an appropriate recipient of the public’s contempt. And the ongoing anger and pain felt by the victims’ families is beyond reproach. What is problematic, however, are government officials opportunistically feigning indignation and perpetuating ignorance about a system operating the way it was designed to.
Before addressing the trouble with these types of reactions by lawmakers, it is useful to briefly outline the basis on which classification decisions are made.
Federal prisons are operated by the Correctional Service of Canada (CSC) and fall within the purview of the Public Safety Canada. Their statutory foundation is the Corrections and Conditional Release Act (CCRA), as well as the accompanying Corrections and Conditional Release Regulations (CCRR) and a set of administrative rules known as Commissioner’s Directives (CD). All aspects of carceral life are impacted by these three sources of law and policy.
Section 3 of the CCRA states that “the purpose of the federal correctional system is to contribute to the maintenance of a just, peaceful and safe society,” and purports to accomplish this by:
(a) carrying out sentences imposed by courts through the safe and humane custody and supervision of offenders; and
(b) assisting the rehabilitation of offenders and their reintegration into the community as law-abiding citizens through the provision of programs in penitentiaries and in the community.
Although s. 4(c) of the CCRA states that “the Service uses the least restrictive measures consistent with the protection of society, staff members and offenders,” s. 3.1 makes it clear that “the protection of society is the paramount consideration for the Service in the corrections process.”
A prisoner’s security rating is classified as being maximum, medium or minimum. The classification process is multifaceted, but is guided in significant part by the relevant considerations set out in s. 17 of the CCRR, including (but not limited to) the seriousness of the offence committed, the prisoner’s performance and behaviour under sentence, their social and criminal history, whether they are designated as a dangerous offender, the presence of mental illness or disorder and their potential for violence.
Section 18 of the CCRR and CD 705-7 (titled Security Classification and Penitentiary Placement) provide that a prisoner shall be classified as maximum security where the prisoner is assessed by CSC as “presenting a high probability of escape and a high risk to the safety of the public in the event of escape,” or as “requiring a high degree of supervision and control within the penitentiary.” In contrast, a prisoner is classified as medium security where the prisoner is assessed by CSC as “presenting a low to moderate probability of escape and a moderate risk to the safety of the public in the event of escape,” or as “requiring a moderate degree of supervision and control within the penitentiary.” CSC is required to give prisoners reasons for their initial security classification and any subsequent changes to it.
Former Minister Mendicino presumably has a detailed understanding of how the prison system works and how classification decisions are made. The former minister knows (or ought to know) that it is not uncommon for a prisoner — even one convicted of terrible crimes — to cascade from one security classification to another after nearly three decades in custody. The legislated purpose of security classifications is not one of retribution but of administration. Furthermore, classification to a medium security prison does not mean that a person serving a life sentence will be released imminently or at all.
When the former minister and others ignore or obfuscate such information, we lose an important opportunity for straightforward public education and engagement on these and other aspects of Canada’s prison system.
When the former minister and others ignore or obfuscate such information, we lose an important opportunity for straightforward public education and engagement on these and other aspects of Canada’s prison system. But the practice of prioritizing a speculative qualm with classification over known classification issues of a systemic nature is long-standing. For example, over the last decade, the federal prison ombudsman, which is the Office of the Correctional Investigator (OCI), has repeatedly made findings that even though Black prisoners are on average less likely to reoffend, they remain “overrepresented in maximum security and underrepresented in minimum security.”
They are also confined to maximum security for longer periods than other demographic groups. This significantly and adversely impacts access to rehabilitative correctional programming, educational opportunities and institutional employment. Furthermore, the OCI has also found that “[t]hough only one in five Black persons had an identified gang affiliation, discriminatory and prejudicial attitudes by some CSC staff often meant that those without a gang affiliation were labelled and treated as such.” The auditor general of Canada and the Standing Senate Committee on Human Rights have made similar findings. Notwithstanding this, no meaningful reforms have been implemented by Public Safety Canada, even though the OCI has pointedly stated that “CSC must systematically examine security classification at admission to minimize unconscious bias, discriminatory practices, and systemic barriers (e.g., access to correctional programming, frequent changes in POs, and long periods without POs) for Black individuals seeking to cascade.”
It is not uncommon for government officials to amplify political rhetoric about prisons over a meaningful discussion of the legal framework that determines how and why they operate as they do. Nor is the practice inadvertent. In fact, it serves a number of purposes, such as cultivating an appearance of responsiveness, quelling public discontent and curbing critical thought and dialogue on controversial issues. While this may yield benefits for the government of the day, it is ultimately detrimental to a public left uninformed about how carceral systems work. This is aggravated further when rhetoric extends dangerously beyond media soapboxes to shape prison law and policy in a manner that is reactionary rather than evidence-based.
What should concern the public is when government officials act expeditiously on the basis of speculation to interfere in classification decisions about a single notorious prisoner (ostensibly one with no realistic chance of release), but concomitantly refuse to take action on the well-established over-securitization and systemic discrimination impacting wide segments of the prison population. This approach to carceral issues makes us none the safer. To the contrary, it maintains a false sense of security among the public and reinforces the disadvantage faced by many prisoners who will one day be released and reintegrate into society. This is the price that we pay for such political posturing. And it is one that we will ultimately learn we cannot afford.
Shane Martínez is a criminal defence and human rights lawyer at Martínez Law and an adjunct professor of Prison Law, Policy & Reform at Osgoode Hall Law School. He can be found on Twitter @martinezdefence .
Is Virtual Justice Real Justice?
Originally published on Law360 Canada, formerly, The Lawyer’s Daily, © LexisNexis Canada Inc
When the pandemic first hit in the spring of 2020, I would not have believed that our courts could do so much virtually as has been done. Indeed, what was made possible through video platforms is simply astounding. By and large, the wheels of justice continued to turn much better than we could have anticipated.
But now that the courts have fully “re-opened” and that in-person hearings are again possible with very few if any restrictions, an important question emerges: to what extent should parties and counsel be required (or at least entitled) to appear in-person for court proceedings henceforth?
Judges and advocates are quite divided about the benefits and costs of “virtual justice,” namely the ability to appear in court through sophisticated software without actually being in court. Some believe that platforms like Zoom and Teams are actually better than in-person attendance. In their view, the ability to communicate with the court virtually is essentially equal to that of in-person attendances, while access to justice is enhanced by eliminating the need to travel to the court. Judges, parties and their counsel, and witnesses are able to participate effectively from their offices or their homes, thus making justice cheaper and more convenient.
Even the Supreme Court of Canada seems to have subscribed to this view, as it has now extended its directive that parties granted the right to intervene cannot appear in-person, but must do so virtually. A good argument is a good argument, it is said, and it makes no difference whether it is delivered in person or remotely.
Thus, many argue that the world of in-person attendances is old and passé, and that in-person hearings should become the exception rather than the other way around; that the convenience and increased access to justice made possible by virtual hearings vastly outweigh any potential benefits in-person hearings might offer. With respect, I disagree. For the following reasons, I hold that, except for the most routine matters (such as case management conferences designed to address timetables) or in exceptional circumstances (e.g., limited evidence from a far-away witness), that view is unsound.
So, why do we have in-person hearings?
THE QUALITY OF JUSTICE
One major reason for insisting on an in-person hearing is that in-person attendance promotes communication. The court can observe witnesses directly; examinations and cross-examinations are more natural (and often more effective) when a witness “feels” the court: there is undoubtedly a “disciplining” effect on all the participants being in the same room in the presence of a judge. Moreover, exchanges between the bench and counsel are unquestionably more natural and dynamic when everyone is together in a courtroom: there are no screens to “freeze” and participants that forget to “unmute” themselves.
It may well be that some witnesses’ evidence will not be affected by the medium of delivery, be it virtual or in-person. But no one will ever convince me that the discipline that is introduced by a witness being physically in a courtroom, with a judge looking over his or her shoulder, is not real and preferable to the witness who is sitting in his living room, cup of coffee to the side and slippers on his feet. You can never know in advance whether a witness’s evidence might be affected by the fact that she is not in the presence of a judge.
As for exchanges between counsel and the bench, whether in the context of significant motions, arguments during trials, or appeals, there is a dynamism that is introduced by in-person hearings that is simply absent in virtual hearings. Remote hearings are just that: they are remote, and I for one never feel the same sense of immediacy as I do when I can look at a judge directly, as opposed to being limited to the view that is offered on my small screen by whomever is managing the remote hearing.
There are many definitions of “virtual.” We can use it in the technical sense of using a computer to approximate a “real” experience. The question is whether this approximation suffices. It may well be that, in many cases, this approximation does suffice, but the fact is that we can never be certain in which cases that is so. Yet as long as we know that we are not actually in a real courtroom, we also know that things could have turned out differently had we been there in person.
THE INSTITUTIONAL VALUE OF IN-PERSON HEARINGS
No one will ever be able to demonstrate that the quality of justice is detrimentally affected by remote hearings. But short of the access to justice arguments, which I address below, I don’t believe it can credibly be argued that virtual hearings are qualitatively better than in-person hearings. At best, they are rough equivalents. But why should a party that has taken a dispute to court ever be left in any doubt that the result might have been different if the witness being cross-examined by her counsel had been sitting 10 feet from the judge, rather than viewed through a small screen and sitting 500 kilometres away? Or if her counsel had been able to engage in the dynamic cut and thrust of legal argument with the judge, which, on screen, is virtually impossible (pun intended)?
In life, all of our most important activities are almost always carried out in person. You can fire a person by e-mail or on Zoom if you want, but is that not the height of hypocrisy or cowardice? You have selected such an impersonal mode of communication because you wanted to avoid the personal contact that being in the same office produces, which proves that remote communication is lacking the human immediacy we prize. We can now attend weddings and funerals online. But will anyone seriously believe that doing so is as meaningful to the newlyweds or the grieving as if you had made the effort to attend in person?
For almost anyone who takes a case to trial or on appeal, the matters at issue are extremely important. Between the time parties file the court papers and the ultimate decision being handed down, the hearing on the merits is often the only time that parties actually see their judge(s) in action: why should they ever be denied the right to appear in person and be assured that the result was not affected by the mode of hearing chosen by the court? And why should judges not want that same assurance? For, in reality, no one can know that the evidence or the arguments might not have been any different in an in-person as opposed to a virtual hearing. The institutional value of justice depends on it being delivered optimally, and that in turn requires that citizens not second-guess its mode of delivery.
THE ACCESS TO JUSTICE ARGUMENT
The main argument in favour of virtual hearings is that they promote cost efficiency and access to justice. By and large, I believe this argument is misconceived. Obviously, for routine matters, it is not (and never has been, or at least not since the telephone has been invented) justified to require in-person attendances.
But for matters that actually do go to trial or on appeal, the cost savings involved in proceeding remotely will usually be minuscule compared to the cost of preparing for and arguing the case (remotely or in-person) itself. Let us remember that only a very small proportion of cases launched ever go to trial, and fewer still go on appeal. Those that proceed obviously have some importance for at least one of the parties involved. Most cases will be heard in a courtroom relatively proximate to the lawyers and parties involved. In any event, the cost of travelling to court compared to the hundreds of hours that will be involved in preparing and arguing the case will be insignificant.
The same is true of cases at the Supreme Court of Canada — including interventions where the court only allows five minutes. On the face of it, it does seem ridiculous that a party should want to travel to Ottawa to deliver a five-minute argument. I will not renew here my criticism of that five-minute rule, but I say that if an intervenor does want to send counsel to be in court in person, it should be their right to do so. For paying clients, the costs of preparing an intervention vastly outweigh any travelling costs; for pro bono matters, such costs will typically be absorbed by the lawyer or firm representing the client, and therefore do not constitute an impediment to access to the court. In either case, therefore, the case for denying access to the court to any party or intervenor wishing it is, with respect, ill-founded.
CONVENIENCE IS NOT SYNONYMOUS WITH JUSTICE
Lawyers who prefer in-person appearances do not, as is sometimes alleged, do so for selfish economic reasons. In fact, many law firms have benefited from the institution of remote hearings and client meetings as they have been able to reduce their operating costs significantly. And, truth be told, it is often much more convenient to attend a hearing from one’s home — sometimes even from one’s cottage — and dress just enough so that the camera reveals that you are properly attired even if you are wearing jeans and sneakers under your robe.
But impressive courtrooms and proper court attire exist because they signify that the delivery of justice — like a vote in Parliament or a church service — are among the most important human activities we undertake.
But impressive courtrooms and proper court attire exist because they signify that the delivery of justice — like a vote in Parliament or a church service — are among the most important human activities we undertake. We could, no doubt, dispense with all our court buildings, and allow judges to sit behind a virtual background made of historical photos of their former courthouses, and if they need offices at all, rent space in some cheap mall wherever they live. Why do we need fancy and very expensive courtroom buildings to maintain and restore at huge public expense if virtual justice is just as good (if not better as some contend) as the old-fashioned kind delivered in person? Governments, for whom the justice portfolio seems increasingly to be of little importance, would no doubt relish the prospect of transferring all judicial proceedings online under the pretext that its quality is in no way diminished and that doing so promotes access to justice.
But by making things too easy, don’t we risk diluting and debasing the importance that we have attached — and should continue to attach — to the delivery of justice? Why did we erect those impressive buildings in the first place, be it the Supreme Court building in Ottawa or the historic courthouses in many of our smaller towns and cities?
Why are we spending hundreds of millions on restoring the House of Commons, the Senate and the Supreme Court of Canada if virtual is as good as in-person attendance? Why did we insist that lawyers and judges dress formally — sometimes with judicial robes and gowns — when they undertook their tasks in the justice system? Why did we assume that almost all significant proceedings need be carried out in person rather than in writing or by phone?
I submit that, while promoting access to justice is undeniably important, it should not be confused with informality and mere convenience if the price is to debase the exercise such that it becomes equivalent to an online commercial transaction. Most of us still dress up and prefer to attend weddings and funerals in person because, while that requires effort and to that extent is less convenient than wearing casual clothes and watching the proceeding on our computer screen, the effort and costs attest to the importance that we attach to the proceedings.
There is little doubt that access to justice remains a very significant and unresolved problem in this country (as in many others, like the United States and the United Kingdom). But, in reality, it will not in any meaningful way be addressed by insisting that significant matters before the courts should as a rule be heard remotely. Not only would the costs savings involved be miniscule, but whatever convenience for the judiciary and the bar would come with the certain cost in which a losing party that had preferred an in-person hearing would always be left with the doubt that the outcome might have been different if the matter had proceeded in-person, and ultimately with the cost that our justice system — which is already undervalued — would be debased further.
Virtual reality is great, but it is not reality. Nor is virtual justice the real justice we should aspire to.
Guy Pratte is senior counsel at BLG . He is a former president of The Advocates’ Society and a fellow of the American College of Trial Lawyers.