Ian Kerr, who passed away far too young in 2019, was an incisive scholar and a much treasured colleague. The wit that sparkled in his papers was matched only by his warmth toward his friends, of whom there were many. He and his many co-authors wrote with deep insight and an equally deep humanity about copyright, artificial intelligence, privacy, torts, and much much more.
Ian was also a valued contributor to the Jotwell Technology Law section. His reviews here display the same playful generosity that characterized everything else he did. In tribute to his memory, we are publishing a memorial symposium in his honor. This symposium consists of short reviews of a selection of Ian’s scholarship, written by a range of scholars who are grateful for his many contributions, both on and off the page.
Table of Contents
Ann Bartow on The Implications of Digital Rights Management for Privacy and Freedom of Expression
Ryan Calo on Emanations, Snoop Dogs, and Reasonable Expectations of Privacy
Kate Darling on Bots, Babes and the Californication of Commerce
Michael Geist on If Left to Their Own Devices…: How DRM and Anti-Circumvention Laws Can Be Used to Hack Privacy
James Grimmelmann on The Death of the AI Author
Eldar Haber on Schrödinger’s Robot: Privacy in Uncertain States
Woodrow Hartzog on Personal Relationships in the Year 2000: Me and My ISP
Mireille Hildebrandt on Prediction, Pre-emption, Presumption: The Path of Law After the Computational Turn
Margot Kaminski on The Devil Is in the Defaults and Soft Surveillance, Hard Consent
Kate Klonick on Chief Justice John Roberts is a Robot
Karen Levy on Digital Locks and the Automation of Virtue
Kristen Thomasen on Pre-Natal Fictions and Post-Partum Actions
Jotwell Reviews by Ian
Juridical Delusions of Control? (July 2010)
Auto-Reportage and the Enlightened User (Nov. 2012)
Ian Kerr and Jane Bailey, The Implications of Digital Rights Management for Privacy and Freedom of Expression, 2 Journal of Information, Communication & Ethics in Society 87 (2004)
Evaluating legal scholarship from 2004, especially articles that focus on technology, can be a freighted process. Only the most astute predictions about the future hold up fifteen years after publication. But Ian Kerr was an unusually forward-looking scholar. He was highly motivated to use law as a tool to improve the interface between technology and humans. Kerr and his co-author Jane Bailey demonstrated exceptional perspicacity within the relatively modest length of this exceptional article. It continues to resonate in 2019.
Kerr and Bailey frame Digital Rights Management (DRM) as a new form of social control, and one that is particularly perfidious for public institutions such as libraries. They focus on technological protection mechanisms, and associated legal measures which legislatures have enacted to create enforceable rights for entities that employ DRM technologies against consumers. Looking specifically at the surveillance function of DRM and its ability to unbundle copyrights into “discrete and custom-made products,” Kerr and Bailey warn that DRM “has the potential to seriously undermine our fundamental public commitments to personal privacy and freedom of expression.”
Kerr and Bailey next focus on freedom of expression, identifying the freedom to speak and the freedom to access information as the freedoms most at risk from copyright-driven DRM. Freedom from government restrictions, they observe, may not be the appropriate goal. Instead governments may need to act to preserve freedom of expression in the face of behavior by private actors. Finally, they conclude that conscious choices could result in DRM that improves consumer access to privacy-friendly content, rather than giving increased control to content owners that will favor corporate goals over the public interest.
There are many reasons to love this article. It is smart, readable, and lays out complicated issues in a very accessible way. It describes a fundamental conflict between DRM and personal privacy that was extant in 2004 and remains critical, under-recognized, and poorly addressed today. It is but one example of Ian Kerr’s scholarship that has remained very relevant for a long time, and will continue to be useful going forward. It reminds readers of the fresh and exuberant manner in which Ian expressed himself. It is also an example of his many successful collaborations with his colleagues. His brilliant insights will continue to educate and inspire for many years to come.
Ian Kerr and Jena McGill, Emanations, Snoop Dogs and Reasonable Expectation of Privacy, 52 Criminal Law Quarterly 392 (2007)
Ian Kerr was a man who loved words. A court’s logic is not “flawed”; it is “pickwickian.” New technologies locate, track, process, and store. But they also melt, pawn, unzip, curl, and coat.
Rereading Emanations, Snoop Dogs, and Reasonable Expectations of Privacy, which Ian co-authored with his student Jena McGill in 2007, I marveled anew at the sheer literary chutzpah. Here is a law review article that opens with an intellectual history of emanations dating back to ancient Greece and ends with Sting lyrics. And yet the argument winds up being cited approvingly by the Supreme Court of Canada the next year. Sometimes Ian’s work reads like legal scholarship from an alternate universe of the better-read.
Emanations, Snoop Dogs, and Reasonable Expectations of Privacy is concerned with the direction of Canadian privacy law in the wake of Regina v. Tessling—Canada’s version of Kyllo v. United States—in which the Canadian Supreme Court upheld a warrantless heat scan of a house on the basis that the defendant’s subjective expectation of privacy in the distribution of heat patterns across the exterior of his home was not objectively reasonable. Some lower courts had begun to extend the logic of Tessling to dog sniffing cases by reasoning that smells, too, are outside the container and hence fair game to law enforcement. Ian and (today) Professor McGill’s purpose in the paper was to critique this extension before it becomes the law of the land.
The authors offer several arguments that Tessling is inapposite to dog-sniffing; one in particular stands out. Extending Tessling in this way ignores the inherently normative dimension of police action. The use of thermal imagining on a home that an informant has already flagged to police as a marijuana nursery has a vastly different social significance than, for example, bringing a police dog into a school to randomly sniff students’ lockers and backpacks. Analogizing these situations is only possible by stripping away the social context and reducing the interaction to a desiccated scientific account of “inside” versus “outside,” and “internal” versus “external” sources of information.
In pressing their case, Ian and McGill invoke Oliver Wendell Holmes, Jr.—courts that ignore the inherently normative dimension to privacy violations behave like Holmes’ bad man, who famously conceives of law as a mere prediction of what courts will do in practice. This move is powerful, and classic Ian: the seamless vacillation between doctrine and jurisprudence in service of a deeply humanistic thesis. To Ian, theory was practice, and both aim squarely at human flourishing.
Other aspects of the article exemplify Ian: That he wrote the paper with a student he was in a position to support and empower. That the article anticipates technological developments, from EEGs to fMRIs to machine learning, that have the potential to amplify the harms of today’s poorly thought-out legal constructs. That the conclusion again sidesteps convention and, rather than summarize the argument as most articles do, introduces a note of wisdom and humility.“This ending,” they write, “is really just a beginning.”
It is hard for me to articulate how much Ian and his scholarship have meant to me. Convention has a weight to it. Writing about odd or marginal topics, or writing about mainstream topics in odd or marginal ways, can feel professionally suicidal. But Ian did not feel this weight, or else had the strength to shrug it off. And in doing so, he opened a path for others. You spend four or seven or more years prostrate to the higher mind. But you get your papers, and you are free.
Ian Kerr, Bots, Babes and the Californication of Commerce, 1 University of Ottawa Law and Technology Journal 285 (2004)
2004 was fifteen (15!) years ago. It’s important to spell that out for those of us who feel like the early 2000s were just yesterday. Fifteen years ago, long before anyone had an iPhone, and years before Twitter was founded, Ian Kerr published Bots, Babes and the Californication of Commerce about the use of virtual agents in commerce. In this piece, he predicts a web of bots that can garner people’s trust and exploit them on behalf of corporate entities.
At a time when forward-thinking scholars were delving into the issues of electronic contract intermediaries, Ian adds a foreshadowing element that few others recognized: social manipulation through virtual agents. This is an idea that is still, 15 years later, increasingly important. Now, online bots are rumored to influence elections, raising concerns about manipulation and trust. Virtual assistants are piling into people’s homes and lives. Amazon has entire teams devoted to Alexa’s “personality,” including what beer she likes (which differs depending on the country). And this is only the beginning. Ian was truly visionary.
Ian argues that a friendship-like trust in bots can be exploited in ways that go beyond our contemporary thinking on misrepresentation, undue influence, and breach of privacy. Because the companies behind the bots can socially persuade people to (“consensually”) reveal personal data and more, he calls on policymakers to reevaluate their consumer-protection principles and study automated manipulation in automatic electronic commerce.
Drawing on philosophers such as Aristotle and Descartes, Ian puts these persuasive bots in the larger context of over a millennium of automation, weaving together vivid examples and adeptly bringing in new (at the time) work from the field of human-computer interaction (HCI), like affective computing. The interdisciplinary approach of this paper is a great illustration of the important questions at the intersections of fields, which few scholars are able to tackle in such depth.
Even the outdated details of the paper, like an example of buying a Palm Tungsten T2 online, or a flirtation with a company rep bot named Nicole, end up highlighting through contrast how well his main points hold up over time.
A note on style: Ian titles, begins, and ends this piece with apt lyrics by the Red Hot Chili Peppers. His work is sprinkled with examples designed to make the reader smile. Around the time I met Ian, my advisors were discouraging me from trying to make my material accessible and entertaining, concerned that nobody would take me seriously. But Ian’s style quickly became my gold standard. While Ian was wonderful about not taking himself too seriously, nobody can deny that his work is serious. His scholarship is excellent, it is prescient, it is rigorous, and it is a delight to read. He gave people like me license to be creative in our writing, and to speak to audiences beyond our colleagues. And his legacy challenges all of us to be so bold as to freely and unapologetically spread joy in all of our academic endeavors.
. ·.·.><((((º> . ·.·.><((((º> . ·.·.><((((º> . ·.·.><((((º>
Ian Kerr, If Left to Their Own Devices…: How DRM and Anti-Circumvention Laws Can Be Used to Hack Privacy, in In the Public Interest: The Future of Canadian Copyright Law (Michael Geist, ed. 2005)
In Canada in the spring of 2005, the then-Liberal government introduced the first major copyright reform bill in nearly a decade. The bill touched on virtually every aspect of Canadian copyright, introducing novel – and controversial – anti-circumvention rules designed to provide legal protections for technological protection measures (TPMs). Within days of the bill’s introduction, I approached a leading Canadian legal publisher with the idea of publishing an edited, peer reviewed volume examining the implications and effects of the bill. The entire project – drafting, peer review, and publication – would need to be completed within four months to ensure that the resulting work would influence the legislative process. I would need to convince Canada’s leading academics to immediately shelve their summer research plans and instead contribute to the project by submitting their chapters within six weeks. While the publisher had its doubts, I had a secret weapon: Ian Kerr. Ian was one of the first to commit to the project, creating the much-needed momentum that would ultimately see 18 contributors sign on.
The significance of this piece is far greater than its role as the catalyst for a successful book project. At a time when politicians and policymakers largely dismissed the negative implications of anti-circumvention legislation, Ian emphasized the connection between digital copyright rules and privacy protection. His warnings read like the current checklist of the fears associated with surveillance technologies, including excessive monitoring, the “piracy” of personal information, and the need to protect consumers from the use of TPMs and online contracts as “privacy circumvention devices.”
This article is Ian Kerr at the very top of his craft. His ingenious use of language frequently used in copyright discourse – piracy, circumvention – to instead describe risks to privacy jolts the reader and invites a re-examination of the broader implications of anti-circumvention legislation. His ability to see beyond the technologies of today to those of tomorrow is readily apparent, as he in 2005 describes the risk of consumer monitoring of “browsing, sampling or shopping.” His insistence that privacy concerns not be treated as a secondary issue to intellectual property protections foreshadows how he would emerge as one of the world’s leading experts and advocates for robust, effective privacy rules. And his creative policy solutions, including expressly calling for safeguards against privacy circumvention, illustrated his uncanny ability to connect legal theory to public policy.
Most of all, it was Ian’s willingness to speak truth to power that shines through in a chapter that has aged exceptionally well. After Ian persuasively demonstrated the privacy risks associated with the government’s proposed policy, he lowered the boom: “Not a single word, let alone appropriate counter-measures, has been contemplated in connection with the implications of DRM for privacy. Not one word.”
It would take the Canadian government seven more years to enact anti-circumvention rules within the Copyright Act. That law included a specific exception permitting circumvention for personal information protection purposes.
Carys Craig and Ian Kerr, The Death of the AI Author (2019),
I wrote There’s No Such Thing as a Computer-Authored Work to tell copyright scholars they should stop speculating about computer authorship. Treating AIs as “authors” is a solution to a problem copyright law doesn’t actually have, because computer-generated works are not a category distinguishable from more traditional human-generated ones. But I was wrong. Not about AIs or about copyright, but about copyright scholarship. Carys Craig and Ian Kerr’s The Death of the AI Author, a deft blend of copyright theory and literary theory, proves by example that this is a vein worth mining. If you worry about whether AIs lack the “individualized self” necessary for true authorship, they argue, just wait until you hear about humans!
It by now a commonplace that copyright uses a conception of “romantic authorship” to justify and to articulate its choices about when and how to give certain people exclusive rights over certain information: an author is an autonomous and independent figure who creates ex nihilo. It is also common to observe that this conception is ideological: this description of human creativity bears little relation to how actual humans create. They build on each others’ work; they are in constant dialogue with audiences and with other authors; their works are intelligible only against this thickly relational backdrop.
Craig and Kerr take this idea in a thrilling new direction: the same critique applies equally well to AI “authors” and to human “authors.” The computer program that “paints” abstract images really is missing something essential for us to attribute authorship to it. But that something essential is not a physical brain or an ineffable soul: it is the web of social relationships in which humans participate “in a dialogic exchange of meaning.” To be human, in the relevant sense, is not to be biologically homo sapiens, it is to engage in those relationships and that dialogic exchange. AIs (at least, I would add, the kind of AIs we have today) do not so engage. Treating an AI as an author is like propping up a cardboard cutout and introducing it to your friends: no matter how photorealistic it is, it can’t play the roles expected of it. Or, to quote a beautiful passage from Craig and Kerr’s conclusion:
To say authorship is human, that it is fundamentally connected with humanness, is not to invoke the romantic author, and nor is it to impose a kind of chauvinism that privileges human-produced artifacts over those that are machine-made. Rather, it is to say that human communication is the very point of authorship as a social practice; indeed, as a condition of life. As such, we do not think we are being at all romantic when we say: authorship is properly the preserve of the human.
Ian Kerr, Schrödinger’s Robot: Privacy in Uncertain States, 20 Theoretical Inquires in Law 123 (2019)
Can privacy be violated if no human is involved? When robots and AI systems process intimate data about an individual, it may seem that there cannot be a privacy violation, as robots and AIs lack human sentience or cognition. On this view, there is no loss of privacy as no other human gains epistemic access to that individual. But what are the implications for embracing such an all-or-nothing approach in an era driven by code?
In Schrödinger’s Robot: Privacy in Uncertain States, Ian Kerr shows why an all-or-nothing account of robots and privacy is inaccurate. He argues that robots and AI can diminish privacy without sentience, consciousness, or cognition, as they can form reliable beliefs and observational knowledge (which Ian names as “truth-promoting beliefs “) that affect our life chances and opportunities, all without human intervention, oversight, knowledge, or awareness. In turn, these truth-promoting beliefs can implicate privacy concerns.
As the world becomes more reliant on sophisticated artificial cognizers acting with substantial autonomy, the life chances and opportunities of people will become increasingly affected by their truth-promoting beliefs, and hence the risk to their privacy will accordingly increase. Using Schrödinger’s famous thought experiment as an illustration, Kerr argues that even if humans do not directly know what the robot inside a metaphorical box “knows,” privacy could still be violated when an artificial cognizer can act from inside the box in ways that affect the outside world.
As always, Kerr’s work on privacy is timely, important, and thought-provoking. Simply put, it is one of the best pieces of academic literature I have had the privilege to read. Schrödinger’s Robot rightfully challenges the concept of relational privacy, or the need for an “other” in a privacy relationship. It sheds light on the potential implications that arise from the use of robots and AI—not within a futuristic scenario, where robots reach a point of superintelligence—but rather within today’s technological capabilities. It is important for the potential recalibration of protecting human rights, like the right to privacy, in a rapidly changing and challenging digital era.
Ian Kerr, Personal Relationships in the Year 2000: Me and My ISP, in Personal Relationships of Dependence and Interdependence in Law (2002)
Ian Kerr often identified concepts, problems, and solutions years before the rest of the world took notice. For example, his insights into bot anthropomorphization and predictive analytics were prescient. Ian also explored the idea of “information fiduciaries”—information relationships of trust—years before it gained steam in international policy discussions.
In his 2002 article Personal Relationships in the Year 2000: Me and My ISP, Ian proposed looking to the concept of fiduciaries to govern the relationships of dependence between Information Service Providers (ISPs) and their users. The concepts of relationships of trust and “information fiduciaries” are getting some attention in information policy circles these days, having been invoked in Congressional testimony, proposed legislation, and regulatory hearings. Jack Balkin’s intervention into the world of information fiduciaries has been quite influential. And a number of scholars, including me and Neil Richards, Ari Waldman, Dan Solove, and others have explored information relationships of trust in some way since the dawn of modern social media and mobile apps.
But Ian called it at the turn of the century. For Ian, an ISP’s power over people stemmed from how dependent people are on ISPs to protect them. Through technological design and market power, ISPs have “fate control” over their users. Much the same can be said today for the relationship between people and nearly every app, website, platform, and digital service, regardless of how much “control” people are gifted in the form of buttons, settings, and switches. Practically speaking, we are forced to rely on platforms. But they, crucially, do not rely upon us in the same way. Ian noted that this lack of interdependence encouraged ISPs to serve their own interests at the expense of users because it is advantageous to do so. Users are dispensable and, as such, will never be on an equal footing with ISPs.
For Ian, the answer to this vulnerability was to look to the law of fiduciaries, which is designed to “protect those who have come to depend upon others.” This article even forecasts critiques of the information fiduciary model, noting its protean nature and narrow applicability. It also dives into the doctrine of fiduciary duties owed to users of ISPs based upon facts or relationship status.
Ian ultimately concluded that ISPs could and should owe users duties of care and loyalty if they invite a user’s trust and dependence. In doing so, Ian anticipated how the then-embryonic notion of cloud computing might deepen users’ dependencies on platforms by further distancing people from their information. Such a move would make ISPs and other platforms “the stewards of personal information and private communications. In such a world, it would seem only reasonable to expect that the management of such information would be carried out in the best interests of users.”
I love this article because it is a lucid recognition of digital information relationships for what they are and why dependent users should be protected. It is a great starting point for anyone hoping to better understand how a classic legal concept is more relevant than ever.
Ian Kerr, Prediction, Pre-emption, Presumption: The Path of Law After the Computational Turn, in Privacy, Due Process and the Computational Turn: The Philosophy of Law Meets the Philosophy of Technology 91 (Mireille Hildebrandt and Katja de Vries, eds. 2013)
Ian had a knack for detecting emerging concerns, and a rare nose for vintage philosophers who turn out to be spot-on for high tech issues. I have always found Peirce’s pragmatist maxim and Holmes’ definition of law as the ‘prophecy of what the courts will decide’ deeply relevant for predictive analytics. But it was Ian who demonstrated why Holmes is not the forefather of computational pre-emption as the term is used right now (e.g. pre-empting people from not recidivising by refusing them parole based on data-driven predictions). It was Ian who explained that it makes a difference who gets to predict whose actions and decisions:
Although he did not offer a comprehensive theory of legal prediction, Holmes taught us that predictions should be understood with reference to the standpoint of everyday people, from their point of view and their sense of purpose. (P. 96.)
I like this piece lots, because it is easier to quote Holmes as forecasting that ‘[F]or the rational study of the law the blackletter man may be the man of the present, but the man of the future is the man of statistics and the master of economics,’ claiming him as the precursor of ‘law and economics’ in either its ‘rational choice theory’ or in its ‘behavioural economics’ version. It is also easier to suggest that Holmes’ other famous dictum ‘the life of the law has not been logic; it has been experience’ fits well with the modern approach to artificial intelligence, which allows machines to learn from their ‘experience’ (referring in this context to the data on which their algorithms have been trained).
Ian, instead, turns the tables on preemptive legal technology, notably when he goes on to explain that current computational prediction is employed by institutions to preempt or manipulate people, turning the Rule of Law inside out. Instead of providing citizens with tools to foresee the consequences of their actions, predictive analytics is wielded by Big Tech (and soon Big Law?) to make citizens transparent while keeping them ignorant of how their anticipated behaviour informs business models and public policy. As we see in the quotation above, that was not what Holmes was after.
Ian Kerr, thus, did not take the easy way out. Why should he? Smart, witty, thinking against the mainstream but never touting complicated intellectualisms, he knew how to make people think: his readers, his audience, his students, his colleagues and friends, policy makers and I am sure also those in the business of making money. It pains me to face the fact that I am using the past tense here, even though I am grateful to have the chance to make his work live on into the future. His untimely death reminds us that in the end, the future is not predictable and cannot be controlled. What can be controlled is not always what matters, and what matters cannot always be controlled.
Ian Kerr, The Devil Is in the Defaults, 4:2 Critical Analysis of Law 91 (2017)
Jennifer Barrigar, Katie Black, Jacquelyn Burkell, and Ian Kerr, Soft Surveillance, Hard Consent, in Lessons From The Identity Trail: Anonymity, Privacy and Identity in a Networked Society(New York: Oxford University Press, 2009)
Back in 2009, Facebook unveiled a control panel that allowed users to choose whether to share a piece of information with “friends,” friends of friends,” or “everyone.” Many hailed this interface as a huge advance for user privacy, a significant step forward for user control. But Ian Kerr was not impressed. As he wrote at the time, “[t]he devil is in the defaults.” While Facebook gave its users the granular ability to choose an audience for their new posts, it also set the sharing default for users’ names, profile pictures, and more to “everyone.” The company knew full well that the vast majority of its users would never exercise the choice to change the default setting.
In The Devil is in The Defaults, which is (technically) a review of Mireille Hildebrandt’s book Smart Technologies and the End(s) of Law, Ian returned to the centrality of defaults to technology law. He wrote, “I have long held that one of the most important lessons of techno-regulation … is the importance of understanding default settings in technological architectures.”
Drawing, as he often did, on a wealth of interdisciplinary references, Ian pointed to the by-now familiar concept of choice architecture—a concept he and his co-authors brought into conversation with privacy law back in 2006 in Soft Surveillance, Hard Consent. It is now commonplace to criticize individual choice and control in privacy law. (Just this month, Charles Warzel in his Privacy Project newsletter bemoaned how “unfair this “personal responsibility” frame seems”). But back in 2006, the idea that companies knowingly fudged and nudged their way into faux consent was far less popular. Consent, Ian and his co-authors counseled then, should be real, ongoing, and continuous under Canadian privacy law. Instead, companies and governments knowingly exploited cognitive biases to get people to “volunteer” personal information they in fact truly valued and wished to protect. Ian and his co-authors offered the example of a Canadian airport notice: “Empowered by self-determination, the passenger can choose between volunteering to be searched and not taking their flight. (Yeah, right.)”
If we don’t have choice, what we do have is default architecture. In The Devil is in the Defaults, Ian used the lens of Hildebrandt’s book to classify a number of kinds of defaults. Where the essay is most effective—and most Ian—is in the closing discussion of the relationship between technological and legal defaults. Any student of law and technology knows that code constrains and governs behavior. But the idea of a factory default, Ian observed here, is not about constraining users, but accommodating their unforeseeably diverse needs. Contract law itself similarly has default rules, which like factory settings “will be employed unless they are intentionally displaced.”
Ian closed the essay by calling for an emphasis specifically on defaults, not just on design. Returning to the Facebook example, Ian called for law to require “that the strictest privacy settings automatically apply once a customer acquires a new product or service.” This proposal is audacious and near brilliant in its simplicity. If we are going to keep harping on about individual choice and control, Ian suggested we flip the, well, default default. Instead of letting users be nudged into making bad “choices,” let’s make it more likely that companies will offer users an effective way to opt out, by having the legal-technological default be something companies would rather users not choose.
Ian Kerr and Carissima Mathen, Chief Justice John Roberts is a Robot (2014)
“Chief Justice John Roberts is a Robot” is not only the engaging title of this playful 2014 article by Ian Kerr and Carissima Mathen, but the premise of its fantastical opening, in which the authors spin a two page tale of the “accidental” rise of a robot to the United State’s highest court. Should we discover that John Roberts was in fact a robot, would what would we make of his legal reasoning? What would we do with his judicial decisions?
It would be a mistake to confuse the unconventional approach with a lack of seriousness. Instead, as the authors meticulously cite, theirs is a thought experiment meant to evoke philosophical questions akin to those of Schrodinger’s cat or Plato’s ring of Gyges. The use of John Roberts, Robot (or “JR-R”) forces readers past what Kerr and Mathen call “traditional caricatures of ‘AI and the law.’” The anthropomorphic use of a robot as a U.S. Supreme Court Justice pushes back in a new and illustrative way on some legal academics’ misplaced ideals in the potential future of “mechanical jurisprudence.” In other words, Kerr and Mathen aim to “provide a reductio ad absurdum of any project that seeks to displace humans from the activity of judging.”
“Chief Justice John Roberts is a Robot” is a fun, easy-to-read piece that thoughtfully explains core philosophical issues that arise whenever discussion of law intersects with the future use of algorithms, artificial intelligence, and robotics. But more than just an incredible road map, it is also a powerful normative argument for the importance of keeping humans in the legal-reasoning loop– not only for what AI cannot bring to bear on such a problem, but for what humanity can. Kerr and Mathen carefully distinguish between the role of humans versus and the role of humanity. The latter is reflected in the core set of morals, values, and norms that we adhere to in forming our communities, and consequently in a constantly moving set of attributes that Kerr and Mathen believe can best be preserved by keeping humans in the legal loop.
At over forty pages, the article is a one-of-a-kind discussion of AI and the law, and also a beautiful representative testament of the invaluable and unique contribution Kerr made to law and technology. As the paper ends, the authors quip that “sometimes the sun comes up even though the party is still going strong” in referencing the many more issues the paper could address but cannot because of their submission deadline. Since its publication, the sun has continued to shine ever more brightly on these important issues. Thankfully, in large part because of Kerr’s incredible work, the foundational fête has been a robust, inclusive, and thoughtful intellectual endeavor.
Ian Kerr, Digital Locks and the Automation of Virtue, in From “Radical Extremism” to “Balanced Copyright”: Canadian Copyright and the Digital Agenda 285 (Michael Geist, ed. 2010)
The first piece of Ian Kerr’s scholarship I ever read was Digital Locks and the Automation of Virtue. I had no idea at the time who Ian Kerr was, but I was bowled over by the clarity of his expression and the depth of his ideas. I remember reading a printout of the piece while standing aboard a crowded subway train, underlining passages enthusiastically while trying not to lose my footing.
Digital Locks takes us on an insightful tour of the social history of locks and keys, from the Gordian knot to the Tower of London, from ancient Egypt to Yiddish proverb. Locks are more than just a simple barrier, Ian argues; locks have keys, and some people have keys while others don’t. Locks, then, can be understood as “technologies of permission”—some of the earliest access-control devices, facilitating different abilities between key-havers and key-lackers.
Like their analog analogues, digital locks—commonly, restrictions on access to and use of copyrighted materials, instantiated in code—also enable differential permissioning. They do so automatically and with precision, often with the added facility of keeping records of access, and backed by anticircumvention rules that penalize lock-breaking. The policy implications of using digital locks as a form of social control are myriad. They impair all sorts of individual freedoms, from privacy rights to the freedom to tinker, from fair use to free expression—but Ian’s particular, and unique, interest in Digital Locks is in how such tools might impede our moral development by attempting to “automate virtue.”
Ian asks: what would the world be like if we used this approach universally—not just to protect copyright, but to prescribe and proscribe conduct across domains and environments? What would it be like if we lived in a world where mistakes were simply not permitted? Drawing on a variety of anecdotes from his own life—and it is a true pleasure to read about young Ian on the impossible-to-crash go-carts at Disneyland, or “schlep[ping] whatever [he] could carry” from his geofence-disabled shopping cart at Loblaws to his car, several hundred meters beyond the property line—Ian explains that “the preemptive nature of digital locks leaves no room for forgiveness.” They automate compliance unflinchingly—leaving no room for mistake, excuse, or disagreement.
Ian’s argument is that people need wiggle room. If you force people to do right, then you’ve missed the point. It’s the choice to be virtuous, when one knows other paths to be possible, that makes a person virtuous. Technology that forces our hand toward right robs us of that choice. In Ian’s words: “the very notion of automating virtue is an oxymoron. The preemption of wrongdoing does not a virtuous person make.” To paraphrase Ian (and Aristotle before him): we become good by doing good, and we cannot do good if we aren’t given choices not to do good.
Further, the “right” path as forged by lock-makers isn’t necessarily the “right” path we might agree on democratically. Rather, it’s the course of action most useful for protecting the lock-maker’s property, monopolistically imposed. In the copyright context, there’s a delicious irony in the juxtaposition of protecting creatively authored works via digital locks—precisely by denying people the means to become moral authors of their own lives.
Digital Locks presages whole paths of thinking in law/tech. Ian’s examples of forced goodness run the gamut from bioengineering to autonomous vehicles—and echoes of his ideas can be found in today’s debates about workplace automation, algorithmic interpretability, and much in between.
When I first read Digital Locks, I hadn’t met Ian Kerr. But the piece is all the more striking these years later, after knowing Ian for an all-too-brief time. The value of the choice to be good wasn’t, for him, merely a clever academic riposte in a policy debate. It was the driving force behind his character. Ian seized every opportunity he could to exercise that choice toward virtue, toward generosity, and toward joy. He chose to be good when he didn’t have to be, and that choice is what mattered to so many.
Ian R Kerr, Pre-Natal Fictions and Post-Partum Actions, 20 Dalhousie Law Review 237 (1997)
This early article of Ian’s builds on his doctoral work on legal fictions. For me, as a doctoral candidate under Ian’s supervision, reviewing this piece was a chance for some IanKerragement (as he called it) – an opportunity to look closely at how my mentor approached his own doctoral research, and gave it practical effect. Ian approaches this paper with his classic, seamless blending of philosophy, logic, law, and most centrally, recognition of human impact. The paper was cited by the Supreme Court of Canada and helped shape a decision that, among other things, safeguards the autonomy and privacy interests of pregnant people in Canada.
The article addresses the legal question of whether a fetus (once born) should be able to sue his or her mother in tort for injuries caused by her negligence while the baby was in utero. This question arose in a Canadian trial decision that allowed a child – injured as a fetus by his mother’s negligent driving – to successfully sue her for damages (Dobson v Dobson). That decision was upheld on appeal, and eventually (after Ian’s article was published) made its way to the Supreme Court of Canada.
Ian had a talent for making the complicated simple. This paper is no exception. He masterfully deconstructs the logic of the trial and appellate decisions, which say that a child could sue for in utero injuries because he was a foreseeable future plaintiff. Ian reminds us that only a person (defined in law) can sue in negligence. And if the fetus at the time of an injury was a future legal person, then logically he was not a legal person at the time of the injury, and therefore cannot sue his mother. It seems pretty straightforward when Ian explains it.
There’s a deeper layer to this paper as well, though. In deconstructing the lower courts’ reasoning in Dobson, Ian also makes philosophical arguments about the use and impact of legal fictions. A legal fiction is, as Ian explains, “a false assumption of fact made by a court as the basis for resolving a legal issue.” One of the central reasons judges use this legal mechanism is “to reconcile a specific legal result with an established rule of law.” (p 240) The use of legal fiction allows the court to preserve an established rule, while also avoiding the application of the rule in that particular circumstance.
In the Dobson decisions, the courts imported a legal fiction from property law, without careful scrutiny, into a tort law setting. The courts use the fiction of a future plaintiff not to advance significant legal principles, but to undermine them, including the Canadian common-law rule that a fetus only becomes a person in law at birth. The consequences of the trial decision, Ian highlights, could be far reaching and would erode pregnant women’s autonomy. Her lifestyle choices, sleep patterns, nutrition, health care,etc. could all become subject to scrutiny through potential future negligence actions. This fiction would have a significant impact on real human lives. In the end, the Supreme Court of Canada reversed these decisions, recognizing the regrettable reality this fiction would create.
Ian’s careful scrutiny of legal fictions is an ongoing theme through much of his later work, including his last piece The Death of the AI Author (co-written with Carys Craig). As a feminist robotics law scholar, I find a lot of IanKerragement in knowing that his early work had a practical effect on preserving the autonomy of pregnant persons.