What Was Equilibrium Like?

Do police need a warrant before pretending to be a child to attract would-be molesters?

It’s been some time now, but the Supreme Court’s decision in R v Mills, 2019 SCC 22, is worth a comment. This is yet another case in which the Court had to address the application of the right, protected by section 8 of the Canadian Charter of Rights and Freedoms, “to be secure against unreasonable search or seizure” to new-ish technologies. These cases often divide the court, but seldom as much as Mills, where Justice Brown signs a leading opinion for himself and Justices Abella and Gascon; Justice Karakatsanis has a concurrence to which the Chief Justice signs on; Justice Moldaver agrees with Justice Brown (whose opinion is thus, in effect, a majority one) and Justice Karakatsanis; and Justice Martin also concurs, but on grounds quite different from her colleagues’.

The issue in Mills had to do with the privacy interests that one might have in one’s online conversations. Justice Brown usefully sets out the facts. The case developed from

a sting conducted by a police officer, who posed online as a 14-year-old girl, with the intent of catching Internet child lurers. Over two months, [Mr.] Mills sent several messages, using Facebook and Hotmail. Eventually, he was arrested in a public park where he had arranged a meeting with the “child”, and was charged … with luring a child via the Internet. The entire operation occurred without prior judicial authorization. Using a screen capture software, the police introduced a record of the emails and messages as evidence at trial. [2-3; paragraph break removed]

Mr. Mills alleged a violation of his section 8 rights, and sought to have the evidence excluded. To decide whether his rights were indeed at stake, the Court must determine whether he had a subjective expectation of privacy in the subject-matter of the alleged search or seizure, and whether this expectation was objective. As usual, it is this last question ― as Justice Brown notes, a “normative question about when Canadians ought to expect privacy, given the applicable considerations” [20; emphasis Justice Brown’s] ― that causes difficulty here.


For Justice Brown, the key to the case is to be found in “the nature of the investigative technique used by police, and the nature of the relationship between the communicants [sic].” [20] In Justice Brown’s view, in light of children’s special vulnerability, “adults cannot reasonably expect privacy online with children they do not know” [23] ― or persons whom they believe to be such children. The relationship here was one between strangers ― since the purported child was in fact a creature of the police ― and the police, obviously, knew this before they started looking at the communications between that “child” and the accused. Unlike in cases where the police intercept or obtain communications between individuals the nature of whose relationship to one another they do not know, they can be certain of breaching no one’s reasonable privacy interests, and so do not need judicial authorization.

Justice Karakatsanis also finds that Mr. Mills could not reasonably expect the messages at issue to remain private. But in her view, that is because the state, acting through the undercover police officer, was the other party to the conversation, and one cannot expect one’s messages to remain private from the party to whom one deliberately sends them. Police officers do not infringe anyone’s privacy rights be speaking to them, even when they are undercover. Conversations via online messaging or email are no different. While the surreptitious recording of what is ostensibly a purely oral conversation makes what was meant to be ephemeral permanent and so has privacy implications, this issue does not arise when writing was the originally chosen medium of communication. Justice Karakatsanis alludes to concerns raised by interveners about police officers posing as persons to whom individuals might want to confide personal information, but decides that they need not be addressed in this case.

Justice Martin takes a different view of the privacy concerns raised by Mr. Mills. For her, the case raises the issue of “whether the state should be permitted to conduct warrantless surveillance of private, electronic communications, or whether that state surveillance should be regulated”. [72] The answer is that “[i]f the police wish to acquire a record of those communications, … such investigative activities must be regulated. The precise nature of such regulation is best left to Parliament”. [72]

Like Justice Karakatsanis, Justice Martin refers to the distinction between transient and permanent communications. But to her mind, the salient point here is that the distinction is being erased:

Online communications are inherently recorded. Where the intrusive technology used to be in the hands of the state, it is now in our back pockets. … [T]he electronic communications in the case at bar constituted both the conversation and the surreptitious electronic recording of that conversation. [86; 93]

Justice Martin endorses the concerns expressed in cases going back to the 1970s that people will, in effect, self-censor if they know that their words may be recorded and publicized. Indeed, Justice Martin worries about government “subjecting the public to surreptitious electronic surveillance on a mass scale”. [103] And since technology makes the existence of recordings inevitable, there need to be robust protections for their privacy. Nor can the state get around these protections by impersonating someone to whom a citizen may wish to speak privately; if it does so, the fact that the recorded words were addressed to the state agent is beside the point. On the contrary, “[t]he ability to fabricate alternative identities has never been more possible [sic] than it is now”, [106] and this reinforces the need for safeguards against the state taking advantage of this ability to elicit private information from citizens.

Justice Martin is also unimpressed by Justice Brown’s approach to the case. She thinks it inconsistent with the usual, content-neutral approach to section 8 cases. More importantly, saying that communications occurring between particular types of people ― such as those between an adult and a child who is a stranger to him or her ― are necessarily excluded from the scope of section 8 is inconsistent with that provision’s text, which guarantees rights to “everyone”, and

seeks to put courts in the business of evaluating the Canadian public’s personal relationships with a view to deciding which among them deserve Charter protection … Judicial (dis)approbation of an accused’s lifestyle has no place in the s[ection] 8 privacy analysis. [110]

Ultimately, however, Justice Martin finds that while the police breached the accused’s Charter right to be secure against unreasonable search and seizure, admitting the evidence they collected would not bring the administration of justice into disrepute. Thus she agrees with her colleagues about the result of the case, if little else. (As previously discussed here by Peter McCormick, Mills will count as a unanimous case in the Supreme Court’s statistics. It is anything but!)


To a striking degree, Mills illustrates Orin Kerr’s “equilibrium-adjustment theory” of constitutional protections against unreasonable search and seizure. In a nutshell, this theory posits that “[w]hen new tools and new practices threaten to expand or contract police power in a significant way” as compared to a (hypothetical) “year-zero” balance, “courts adjust the level of [privacy] protection to try to restore the prior equilibrium”. (480) Here, the tools and practices ― available both to the police and to the citizens (criminal and law-abiding alike) are online communication platforms that combine possible anonymity with the recording of conversations (and so, as just noted, the erasure of the demarcation between the spoken and the written word).

As equilibrium-adjustment theory predicts, Justices Brown, Karakatsanis, and Martin all frame their reasons as means to preserve or restore a balance of privacy that these developments threaten to disrupt. Thus Justice Brown insists that the means used by the police in this case “did not significantly reduce the sphere of privacy enjoyed by Canadians”. [20] For her part, Justice Martin argues that the “Court must identify the privacy interest [previous] cases sought to protect and ensure that it remains protected as the communication environment evolves”. [90] Justice Karakatsanis is perhaps a little less explicit about her own effort at equilibrium adjustment, by her insistence that written communications have not, in the past, been treated in the same way as oral conversations, and should not be so treated now is also in that vein, as is her concern that “[t]he alternative conclusion would significantly and negatively impact police undercover operations, including those conducted electronically”. [52]

But while all the judges in Mills agree on the importance of preserving the balance between privacy and the ability of police to investigate crime, it is not so clear where exactly that equilibrium really was. At equilibrium, was it the case that adults had no privacy rights in their relationships with children who were strangers to them, Justice Brown suggests? Or that the written word was not private in the way the spoke word was, as Justice Karakatsanis argues? Or, on the contrary, that all relationships, regardless of the parties’ status, and all conversations, regardless of the means used to carry them out, were entitled to privacy protections, as Justice Martin suggests?

Justice Brown’s attempt at defining equilibrium does not persuade me. As Steven Penney wrote in his (rather timelier than mine) comment on Mills on the University of Alberta Faculty of Law Blog, it is difficult, if not impossible, to tell when two persons are “strangers” to one another:

What degree of familiarity with an online persona is required before he or she ceases to become a stranger? Is it necessary to have met the person face-to-face in the offline world? Or is it enough to have had prior oral conversations (whether audio-only or video chats) online? And what, if any, degree of identity verification is required?

Meanwhile, the concept of “children” seems perfectly well-defined, in this context, as people under the age of 18, but this clarity might not be all it seems (because people lie about their age), and comes with its own set of problems. Are close-in-age relationships different? What about relationships between two minors? Why, indeed, draw a hard line at 18, especially outside the context of sexual crimes? Alternatively, must investigations of sexual crimes be treated differently than those of, say, terrorism?

One is tempted to suspect (Professor Penney, I think, hints at this too) that Justice Brown was looking for a very narrow basis on which to resolve this case. But the trouble is that an artificially narrow decision in a difficult case risks both being unprincipled and simply kicking the can down the road. Justice Brown’s opinion is in serious jeopardy on both these counts; indeed, I am inclined to declare it guilty on the second, even if a reasonable doubt might exist as to the first.

But what about the disagreement between Justices Karakatsanis and Martin? Perhaps this disagreement is the latest ― and probably not the last ― manifestation of a recurrent problem in cases about the application of section 8 of the Charter, which I described here when commenting on the Supreme Court’s decision in R v Marakah, 2017 SCC 59, [2017] 2 SCR 608, a case that considered the privacy of a text messaging conversation that one of the parties handed over to the police:

A number of legal issues arising out of new technologies, broadly speaking, have to do with the erasure of the once-clear line between the spoken and the written word. The former was (usually) spontaneous and fleeting; the latter (relatively) deliberate and permanent. But electronic communications combine spontaneity and permanence in a way to which many of us are still only getting used and with which the legal system, unsurprisingly, struggles.

In Marakah, a majority of the court (which included Justice Karakatsanis) held that a reasonable expectation of privacy in the conversation existed, so that the police could not look at it without prior authorization, even with the consent of one of the parties. I think that decision could also be understood as an attempt at equilibrium-adjustment, intended to preserve the previously undoubted privacy of the exact content of personal conversations. Professor Penney argues that Justice Karakatsanis now tries attempts to “effectively overturn” Marakah.

But I’m not sure that this is fair, or that the problem of the newfound permanence of conversations is really what is driving the disagreement between Justices Karakatsanis and Martin in Mills. Justice Karakatsanis does not deny that (in Professor Penney’s words) “Marakah … puts automatically recorded text conversations on the same plane as surreptitiously recorded oral ones”. Rather, she wants to hold on to a distinction that, as I see it, Marakah did not foreclose: that between the state obtaining, by whatever means, a conversation to which it was not a party (as it had done in Marakah), and that between the state itself in effect being a party to the conversation. In the latter case, Justice Karakatsanis argues, “[t]he fact that the conversation took place in a written form … does not transform it into a search or seizure”. [45] In other words, for Justice Karakatsanis, as for the Marakah majority, oral and electronic conversations are alike; those that involve a state representative and those that do not are not.

And, so far as this goes, I tentatively think that Justice Karakatsanis is right: considered by itself, the (electronic) conversation between a suspect and an undercover police officer in Mills is not a meaningfully greater intrusion on the suspect’s privacy than a conversation with an in-person informant would have been. Of course,there is a word-for-word record of that conversation. But, as Justice Karakatsanis says, the suspect knows this in advance. It’s all good and well to proclaim that the “reasonable expectation of privacy” standard is normative, not descriptive; it’s about what people ought to expect, not what they actually expect. But the standard cannot be entirely unmoored from the facts and, in particular, from the fact that we know and indeed use tools (such as search functions) that rely on the fact that the words of our conversations are recorded in real time.

That said, there is also a different equilibrium-adjustment issue in Mills, and it is this issue which, I think, really drives the disagreement between Justices Karakatsanis and Martin: namely, whether the ability of the police to engage in online undercover surveillance will radically expand the use of this investigative technique. Justice Martin think that it will, and so wants to forestall this expansion of police power. She envisions “mass scale” surveillance, previously “inconceivable … due to the practical resource constraints of [in-person] undercover police work” [104] and because of the much greater variety of fictitious personas investigators might be able to adopt online. Justice Karakatsanis is skeptical about this (as is Justice Brown). It is very difficult to say who is right here. Justice Martin, I think is speculating about the future prospects of mass surveillance; for now, at least, I don’t see the prospect of pervasive police stings using fake online personas as anything more than a dystopian fantasy, albeit, to be sure, not an entirely implausible one. And, in fairness, Justices Karakatsanis and Brown are speculating too, hoping that this dystopia does not come to pass, or at least that its development, should it begin in the earnest, will be able to be checked then.


So let me finish with a thought on one way to help prevent the dystopian future. It will perhaps seem naïve, but I think it is actually important. There are two ways to reduce the odds of police investigations unduly intruding on citizens’ lives. One is to constrain investigations once they are launched by limiting the use of certain techniques, requiring warrants, etc. Section 8 of the Charter, and equivalent provisions of other constitutional instruments, do this but, as cases such as Mills illustrate, constructing good doctrine in such cases is not easy. The other way to keep police in check is to have fewer crimes for them to investigate in the first place. As Justice Gorsuch, of the US Supreme Court, observed just a couple of days ago in Nieves v Bartlett,

History shows that governments sometimes seek to regulate our lives finely, acutely, thoroughly, and exhaustively. In our own time and place, criminal laws have grown so exuberantly and come to cover so much previously innocent conduct that almost anyone can be arrested for something. (Slip op. 1-2; that’s pp. 24-25 of the PDF) 

And of course, before one can be arrested, one can be investigated. Perhaps the Canadian situation is not as bad as the American one (I don’t know enough to tell) but, if so, we must work to keep it that way. And here, the Charter ― just like its counterparts elsewhere ― is not going to help us. Only sustained political opposition to overcriminalization ― and, ideally, sustainted political support for decriminalization of a great many things currently considered criminal ― will do the trick.

Can’t Take It

Can the police seize a computer (without searching it) if only one of its co-owners consents?

In R v Reeves, 2018 SCC 56, delivered last week, the Supreme Court held that section 8 of the Canadian Charter of Rights and Freedoms, which protects the right not to be subject to unreasonable search and seizure, prevents the police from seizing (even without searching) a computer located in the common area of a home with the consent of only one, but not the other of the home’s occupiers. As in a number of other search-and-seizure cases, it is Justice Karakatsanis who takes the lead in articulating a narrow view the police powers. Unlike in R v Fearon, 2014 SCC 77, [2014] SCR 621 and R v Saeed, 2016 SCC 24, [2016] 1 SCR 518 she carries a strong majority of her colleagues ― all but one, in fact, on this issue ― with her. And unlike in those two cases, I suspect that Justice Karakatsanis’ pro-privacy disposition has not served her well.

The facts of the case are somewhat quirky. Mr. Reeves had shared a home ― and a computer ― with his common-law spouse, Ms. Gravelle. So far, so ordinary. However, following a family violence incident, Mr. Reeves was barred from being at the home without Ms. Gravelle’s consent, which she eventually withdrew. Still, as a matter of property rights, both the home and the computer were still shared between Mr. Reeves and Ms. Gravelle. Ms. Gravelle also informed the authorities that she had previously found child pornography on the computer. A police officer came, and, with Ms. Gravelle’s consent, he took the computer back to the police station ― where it sat, seemingly of no interest to anyone, for four months, despite the Criminal Code requiring such seizures to be reported to a justice of the peace. Eventually, the police finally concocted a warrant application ― which the trial judge later found to be tendentious and deficient to the point of invalidating the warrant ― and searched the computer, duly finding the child pornography, leading to charges against Mr. Reeves, who argued that the evidence was obtained in violation of his Charter rights and should be excluded.

 There were two possible violations of Mr. Reeves’ rights for the Supreme Court to look into. First, the police officer’s entry into and search of the shared home; second, the seizure of the shared computer. (There was no dispute that the lengthy detention of the computer in violation of the Criminal Code and its search pursuant to a warrant that was subsequently invalidated were constitutionally problematic.) However, for the majority, Justice Karakatsanis does not pronounce on the requirements of the Charter with respect to police entry into a shared home with the consent of only one of its occupiers. She finds that the matter is best left for another time, when it will be more fully argued. Justice Moldaver, in a concurring opinion, agrees that now is not the time to dispose of the question ― and proceeds to lay out a detailed case for why the police have a common law power to enter to speak with one co-occupier of a shared home, while insisting that this argument is only tentative.

For the majority, the case turns on the question of the seizure of the computer. This, in turn, divides into two sub-issues: first, whether Mr. Reeves had a “reasonable expectation of privacy” in the machine; and second, whether Ms. Gravelle could obviate or waive his rights by consenting to the seizure anyway.

As Justice Karakatsanis explains, “[t]he reasonable expectation of privacy standard is normative, rather than descriptive” [28] ― it is not really about what the person concerned expected his or her privacy rights to be in the circumstances, but about where the balance between privacy and societal interests (in particular, in the investigation and punishment of offences) ought to be struck. In deciding this question, Justice Karakatsanis insists that “the subject matter of the seizure was the computer, and ultimately the data it contained about Reeves’ usage, including the files he accessed, saved and deleted”. [30] Even though, as the Supreme Court previously held, a separate warrant would be required to actually search the data contained in the computer, “Reeves’ informational privacy interests in the computer data were still implicated” [30] because he lost control of it, including the ability to destroy it. The data computers contain can be “highly private” [34], and thus not only the search, but also the mere seizure of “a personal computer from a home” “presumptively require[s]” “specific, prior judicial authorization”. [35] This is so even when the computer is shared and no one individual can expect absolute privacy when using it.

As for a co-occupier of a home consenting to the police seizing a shared computer, it “cannot nullify” [41] an existing reasonable expectation of privacy: “[t]he decision to share with others does not come at such a high price in a free and democratic society”. [44] Those others can report suspicions to the police, but it does not follow that the police can do anything they (the others, that is) consent to. It is not their (the others’) rights that are at stake, after all, and the fact that they too may have rights or privacy expectations over the same object or space is beside the point. As for what the police can do if a person actually brings an object in which another has a reasonable expectation of privacy to them, that “remains for another day”. [46] 

In the event, the majority, as well as Justice Moldaver, conclude that the Charter breaches in this case are serious enough to warrant excluding the evidence found Mr. Reeves’ computer. Justice Côté, in a concurring opinion, agrees with this outcome ― even though, as I am about to explain, she does not think that the seizure of the computer amounted to a Charter breach at all. (The Supreme Court, which only considers outcomes in its statistics, will triumphantly count Reeves as yet another unanimous decision ― yet as Peter McCormick recently explained here, it is a mistake to do so.)

On the key issue of the case ― the application of section 8 of the Charter to shared spaces and objects ― Justice Côté takes an approach that is the opposite of the majority’s. Unlike Justice Karakatsanis, Justice Côté directly addresses the question of whether police can enter shared spaces with the consent of a single occupier ― and answers in the affirmative, albeit with a possible qualification. Justice Côté writes that “it is not objectively reasonable for a cohabitant … to expect to be able to veto another cohabitant’s decision to allow the police to enter any areas of that home that they share equally”. [112] Such a veto would amount to a negation of “consenting cohabitant’s liberty and autonomy interests with respect to those spaces”. [112] It would also “require the police to identify, locate and obtain the consent of every person who lives in the home, or has any expectation of privacy with respect to common areas of the home”, [114] which is likely to be unduly burdensome at best, if not quite impossible. And applying this approach to shared virtual spaces or objects ― say, a text messaging chain ― would produce similarly perverse consequences. Meanwhile, the search of common areas of a shared home is unlikely to produce intrusions into a person’s deepest secrets.

Justice Côté takes a similar approach to the seizure of a shared computer. While acknowledging that searching such a computer would require prior authorization, she argues that the mere seizure consented by one of the computer’s co-users is not a violation of the other co-users’ rights ― and thus disagrees with the majority on this key point. Again, it is not reasonable to expect that a co-user will not allow the authorities to seize a shared computer, and concluding otherwise would deny the co-user’s  autonomy. The context of co-ownership and joint control influences the scope of one’s reasonable expectation of privacy. Furthermore, Justice Côté stresses the fact that only the seizure of the physical object, not the information it contains, is at issue, and reproaches the majority for conflating the two. She also points out that the majority’s approach may well amount to recognizing an expectation of privacy not just in favour of a co-owner of a computer, but also of, say, a guest who used it at some point in the past. Justice Côté adds that of course a co-owner should be able to take a shared computer to the police ― and letting the police take it is no different.

As noted at the outset, I am inclined to think that Justice Karakatsanis has indeed gone too far in protecting privacy here. Justice Côté is right that the majority conflate interests in maintaining control of a physical object and those in ensuring the privacy of the data that this object contains. And it is true that, by effectively granting a veto to each co-user (and perhaps even a past user ― perhaps the majority would distinguish that case, but it is indeed unclear how, and it’s unfortunate that Justice Karakatsanis doesn’t address the point), the majority compromise the autonomy ― and interfere with legitimate interests ― of other co-users. It would, as Justice Côté says, be odd if such people couldn’t take things of which they have legitimate control to the police ― and no less odd if they could not invite the police to take such things.

At the same time, a couple of points bother me about Justice Côté’s reasons. First, she might be too sanguine about the prospect of police not gaining access to particularly private information in common areas of shared dwellings. This may indeed be a reasonable assumption if the people living together are what in North America are misleadingly called) room-mates, and elsewhere, more accurately, flat-mates, who each have a private room. But if they are spouses, or otherwise family members, the distinction between common and private spaces within the shared home may not be drawn with any clarity. Perhaps this does not matter after all, but Justice Côté would have done well to address this issue.

I also am somewhat puzzled by Justice Côté’s references to the odd circumstance that Mr. Reeves had lost access to the theoretically-shared home and computer that were the objects of the police’s interest here. It’s not quite clear how much this fact matters to Justice Côté’s conclusions. I think it’s not particularly significant in her reasoning regarding police entry into a shared dwelling, but on the seizure issue, Justice Côté explicitly says that it is “relevant” that Mr. Reeves “lacked control [of the computer] as a result of his own actions”. [130] Yet not only was this “result” an indirect and unintended, albeit foreseeable one, but, more importantly, one is left in some doubt about how Justice Côté thinks more ordinary cases, where this “relevant” factor will not be present, ought to be decided.

Ultimately, Reeves might not be a very important case. The one issue it actually decides, whether police can seize shared computers with the consent of one but not all of their users, may not recur all that often, insofar as people increasingly use personal laptops, tablets, or smartphones. I don’t actually know if they do, but I suspect that they might. Perhaps its chief interest is in the trends that it confirms: Justice Karakatsanis’ role as the Supreme Court’s leading pro-privacy voice, and Justice Côté’s as its leading independent thinker. On the whole, the Court needs both, even when they disagree, and even in cases where neither is quite right. 

Ceci est-il une conversation?

The Supreme Court holds we can expect our text messages to remain private, even on other people’s phones

Last week, the Supreme Court released its eagerly-awaited judgment in R v Marakah, 2017 SCC 59, holding that a person had standing to challenge the admissibility of text messages to which he was a party but which the police had seized from another’s cell phone. The Chief Justice wrote the majority opinion, in which Justices Abella, Karakatsanis, and Gascon concurred. Justice Rowe wrote a brief concurrence, raising some concerns about the future implications of the majority opinion, with which he nevertheless agreed. Justice Moldaver, with the agreement of Justice Côté, wrote a fierce, strongly-worded dissent.

Section 8 of the Canadian Charter of Rights and Freedoms provides that “[e]veryone has the right to be secure against unreasonable search or seizure”.  This right applies when a person has an objectively reasonable “expectation of privacy” in the thing or information that is the object of the search or seizure. There is no question that Mr Marakah wanted and expected his exchange of text messages with an accomplice in a weapons-trafficking venture to remain private. But was he entitled to expect that the police would not read these messages on that accomplice’s phone?

The majority and Justice Rowe think that he was. As the Chief Justice put it, the

interconnected web of devices and servers creates an electronic world of digital communication that, in the 21st century, is every bit as real as physical space. The millions of us who text friends, family, and acquaintances may each be viewed as having appropriated a corner of this electronic space for our own purposes. There, we seclude ourselves and convey our private messages, just as we might use a room in a home or an office to talk behind closed doors. [28]

The information exchanged in these nooks and crannies of cyberspace is, potentially, highly private, and indeed “[i]ndividuals may even have an acute privacy interest in the fact of their electronic communications”. [33] Crucially,

this zone of privacy extends beyond one’s own mobile device; it can include the electronic conversations in which one shares private information with others. It is reasonable to expect these private interactions — and not just the contents of a particular cell phone at a particular point in time — to remain private. [37]

The fact that we might not control all the devices through which this information is accessible is not especially important. It is the information exchanged, the conversation, that is the subject of the expectation of privacy, not whatever device might allow one it view it. And even the fact the person with whom one is texting could disclose the fact or the content of the conversation does not allow the state to read it.

Justice Moldaver disagrees. For him, control is a key factor in the analysis. Justice Moldaver writes that “the reasonableness of a person’s expectation of privacy depends on the nature and strength of that person’s connection to the subject matter of the search”, and “[w]here an individual lacks any measure of control, this serves as a compelling indicator that an expectation of personal privacy is unreasonable”. [98] Justice Moldaver gives a number of examples: DNA in one’s body is private, but DNA traces left on, say, the body of a victim of a crime are not; thoughts recorded in a private diary are private, but not those publicly shared online. [116] While control does not require ownership or exclusivity of access, a lack of control means that information is not in a meaningful sense private.

When it comes to conversations, including conversations conducted by text messaging, Justice Moldaver is of the view that one loses control over what one has said once one has said it. What one’s interlocutor’s phone records is “an independent record”, [128] similar to the notes one might make after a spoken conversation, and within the interlocutor’s exclusive control. Evesdropping on an ongoing conversation, or intercepting text messages as they are being sent, violates a reasonable expectation of privacy. By contrast, just as each party to a conservation is free to share a record or recollection of it, and his or her interlocutor can (subject to any applicable privacy legislation) have no reasonable expectation of privacy in that record, so it is also with a “record” of a conversation conducted via text messaging.

Here, as I see it, is one important point of disagreement between the majority and the dissent. Both are ostensibly agreed that what Mr. Marakah had, or lacked, a reasonable expectation of privacy — or, in other words, “the subject matter of the search was Mr. Marakah’s ‘electronic conversation’ with” his accomplice. [17; 111] But it seems to me that while the majority does indeed approach the case as one about the privacy of a conversation, the dissent sees it as being not about a conversation as such, but rather about a record of a conversation. To repeat, Justice Moldaver accepts that “an electronic conversation” would be private; it could not be intercepted without due authorization. But the messages stored in the cell phone of one of the parties to the conversation are not the same thing. They are like the notes one of the interlocutors took. (Hence the title of this, in reference to René Magritte’s notorious The Treachery of Images, a.k.a. Ceci n’est pas une pipe.) As Justice Moldaver suggests, we can expect not to be eavesdropped on, when having a private conversation, but not necessarily that the contents of that conversation will never be revealed to third parties. So the majority decision makes sense in light of how it understood the issue, and the dissent makes sense in light of its author’s different understanding of the case.

But which of them is correct? I personally find this a very difficult question. A number of legal issues arising out of new technologies, broadly speaking, have to do with the erasure of the once-clear line between the spoken and the written word. The former was (usually) spontaneous and fleeting; the latter (relatively) deliberate and permanent. But electronic communications combine spontaneity and permanence in a way to which many of us are still only getting used and with which the legal system, unsurprisingly, struggles. One of my very early posts, for instance, was about a case that concerned an attempt by a university to punish students who ranted about their professor on Facebook. Student rants about a professor are nothing new, but the fact that they were made online rather than over beers left a record for the authorities to look into and to try (unsuccessfully in the event) suppressing. In a different way, the disagreement about the way to characterize text messaging “conversations” — often created in a spontaneous way, as if the parties were together in the same room, but a permanent record for the police to look at later — exemplifies the same set of difficulties. (This might come out most clearly in Justice Rowe’s brief concurrence.) On balance, though, I am inclined to think that Justice Moldaver’s view makes more sense. The idea of a never-finished conversation, to which one is always an ongoing party, and in which one is permanently entitled to expected privacy, which seems implicit in the majority’s approach, doesn’t quite make sense to me. This is a very tentative thought, however, and a minority view, I gather.

Beyond the characterization of “electronic conversations”, the Chief Justice and Justice Moldaver also disagree about the policy implications of the Supreme Court’s decision. In particular, Justice Moldaver worries that police will not be able to access, without a warrant, “electronic conversations” that are voluntarily tendered to them by one of the parties, even when the conversations are themselves crimes, and the parties disclosing them to the police are victims. A person may, for example, receive a threatening text message, and want to show it to police officers, but it is not clear that the police will be entitled to look without judicial authorization. At best, this will complicate the work of the police; at worst, serious crimes will go unpunished. The Chief Justice responds that these difficulties can be dealt with if and when they arise. For his part, Justice Rowe is not so sure, and I take that it is because he ” share[s] the concerns raised by Justice Moldaver as to the consequences of this decision” [89] that he goes to the trouble of writing separately.

A lot, then, remains to be decided. Privacy issues have been consistently difficult for the Supreme Court, or at any rate more consistently divisive than most others. I find these issues difficult too, so I have sympathy for judges on both sides. That the majority wants to be protective of privacy in a way the majority in R v Fearon, 2014 SCC 77, [2014] SCR 621 (which I criticized here) was not is heartening. (Some people on Twitter were wondering how many of the judges had got smartphones in the meantime. A cynical question, perhaps, but I’m not well placed to critcize those who are cynical about judges, am I?) The question now is whether the pendulum has swung too far in the direction of privacy. It might have, but we will have to wait to find out.

Was Scalia Spooky?

Antonin Scalia’s views on snooping, in the 1970s and later

The Globe and Mail‘s Sean Fine is as good a reporter as he is a bad analyst. Both of his qualities ― an impressive ability to find and tell a great story, and an unthinking belief in simplistic ideological classification of judges ― are on full display in his latest article, a fascinating story of how Antonin Scalia, then a professor at the University of Chicago, was commissioned to produce a report on “United States Intelligence Law” for the McDonald Commission, which investigated the RCMP’s espionage activities and whose eventual recommendations led to the creation of CSIS. Mr. Fine contrasts “[t]he report’s scrupulously impartial (for the most part) author” with the judge that he would become; the former, sensitive to privacy rights if also keen to ensure that intelligence agencies can operate effectively; the latter, in Mr. Fine’s telling, brazenly unconcerned with them, and condoning “torture in some circumstances”. But things are more complicated than Mr. Fine lets on.

Before I get to that, I’ll note little anecdote that Mr. Fine passes over, perhaps because this is a bit too inside-baseball for the Globe‘s readers. Mr. Fine explains that it was Peter Russel, who was the director of research for the McDonald Commission, who recommended then-professor Scalia’s hiring ― on the advice of Edward Levi (Scalia’s boss as Attorney-General in Gerald Ford’s administration) and Herbert Wechsler (a distinguished scholar, notably of the “neutral principles” fame). What Mr. Fine does not mention is that prof. Russel’s recommendation (a scan of which is included in the article) noted that Levi and Wechsler ranked Scalia ahead of none other than Robert Bork. (Prof. Russell, by the way, seems to have had a bit of an issue with names in that memo, referring to “Anthony” Scalia and “Richard” Bork.) Ironically, the Reagan administration would later rank Scalia and Bork in the same order when it came to making their appointments to the Supreme Court. Scalia was nominated in 1986, and confirmed by the Senate on a 98-0 vote; Bork was nominated in 1987 and rejected by the Senate after hearings so bitter that his name became a verb, in which his views and record were arguably distorted out of all recognition by Ted Kennedy and the latest recipient of the Medal of Freedom.

And, to get back to my point, this is a bit what Mr. Fine tries to do with the late Justice Scalia, albeit on a much smaller scale. He makes a point of noting that prof. Russell

would … later be appalled by the justice’s support of originalism – a judicial philosophy in which constitutional rights do not evolve over time, but stay rooted in the vision of the Founding Fathers of the United States. “Originalism is absolute nonsense”,

he quotes prof. Russell as saying. And he refers repeatedly to a “2007 speech” Scalia gave in Ottawa, in which “he was more the suspicious-of-too-many-legal-protections conservative”.  But Justice Scalia’s originalism was neither “nonsense” nor all bad for the protection of privacy rights against over-curious governments.

Prof. Russell, Mr. Fine, and those who think like them ― admittedly, a large contingent in Canada ― might just learn a thing or two from the expanding scholarship documenting the presence of originalism in Canada, and in some cases advocating the expansion of this presence. This scholarship includes (but is not limited to) recent articles by Sébastien Grammond and J. Gareth Morley focusing on the Supreme Court’s opinions on the appointment of Justice Nadon and Senate reform; an as-yet-unpublished paper by Asher Honickman, on federalism; Kerri Froc’s work on women’s rights; and the pair of articles that Benjamin Oliphant and I wrote last year. The first of these, which should come out any day now in the Queen’s Law Journal, shows that contrary to popular belief, the Supreme Court has not squarely rejected originalism, least of all what is arguably the dominant form of originalism now, one focused on the original meaning of constitutional texts (rather than their framers’ intentions or expectations). The second, due to come out in the UBC Law Review later this year, shows that, in fact, the Supreme Court resorts to originalist reasoning in a surprising variety of cases. If Prof. Russell is right that “originalism is absolute nonsense”, then not only has the Supreme Court never renounced it, but in fact large swathes of its jurisprudence (and of that of the Judicial Committee of the Privy Council), are nonsensical too.

But more directly relevant to my present topic is our discussion, in the first paper, of the contrast between Justice Scalia’s reasons, for a 5-4 majority of the Supreme Court of the United States, in Kyllo v United States, 533 U.S. 27 (2001), and Justice Binnie’s reasons for the unanimous Supreme Court of Canada in R v Tessling, 2004 SCC 67, [2004] 3 SCR 432. As we explain (actually, the credit here goes to Mr. Oliphant):

The issue, in both cases, was whether the use of a thermal imaging device by the police amounted to a “search” within the meaning, respectively, of the Fourth Amendment to the U.S. Constitution and section 8 of the Charter. In Kyllo, Justice Scalia … found that because information about what went on within the home ― however collected ― would have been secure from search and seizure at the time the Fourth Amendment was passed, the state cannot now invade that sphere of privacy through the use of new technology.

Justice Binnie, writing for a unanimous Supreme Court, disagreed… Justice Binnie rejected the relevance of Kyllo on the basis that it was “predicated on the ‘originalism’ philosophy of Scalia J.,” [61] and because it is not “helpful in the Canadian context to compare the state of technology in 2004 with that which existed at Confederation in 1867, or in 1982 when s. 8 of the Charter was adopted.” [62]

Tessling is an odd hill upon which to make a stand against originalism. Kyllo, which the Court in Tessling refused to follow, did not restrict constitutional meaning to those realities foreseen by the framers, as originalism does according to the “frozen rights” or “dead” constitution caricature frequently encountered in the Canadian literature. It did precisely the opposite. … Indeed, it is not clear to us just what Justice Binnie is actually rejecting in refusing to follow the “originalist” philosophy underlying Kyllo, or in stating that it is unhelpful “to compare the state of technology” in 2004 with what which existed in 1982. The logic of Kyllo was to deny that changes in technology can diminish the scope of constitutional protection over time; there was no “comparison” of technologies, because changes in technology were irrelevant to the interpretive question of what was protected. (25-26; a paragraph break and a reference removed)

We conclude that

In the ultimate result, and despite frequent and nebulous assertions that the Charter must be read in a “large,” “liberal,” and “generous,” manner, Justice Scalia’s originalist philosophy unquestionably resulted in a more general and robust protection for personal privacy than Justice Binnie’s “purposive” approach to interpreting section 8 of the Charter. (27)

Of course, this is not to say that Justice Scalia was always right, on privacy issues or on anything else. Indeed, this does not even prove that originalism is the better approach to constitutional interpretation than whatever it is that the Supreme Court of Canada is doing. But both originalism and Justice Scalia’s legacy are more complex than many Canadians, including Mr. Fine, tend to assume. We owe Mr. Fine for telling us a story that shed more light on the late Justice’s oeuvre. It’s too bad he tried to shoehorn that story into a simplistic ideological framework that is as misleading as it is useless.

Not Such a Simple Thing

A divided Supreme Court expands the powers of search incident to arrest

A couple of weeks ago, the Supreme Court issued a decision, R. v. Saeed, 2016 SCC 24, that was further evidence of its majority’s expansive views of the police’s powers of search incident to arrest ― and trust in judicially developed checklists to prevent the abuse of these powers. Meanwhile, by writing an opinion which, although concurring with the majority in the result, rejected its approach, Justice Karakatsanis confirmed her role as the Court’s leading ― if only in dissent ― privacy-protecting voice. To that extent, the case was a reprise of the Court’s earlier decision in  R. v. Fearon, 2014 SCC 77, [2014] 3 S.C.R. 621, where it considered the powers of the police to search an unlocked cell phone incident to arrest. (I commented on that case here.)

The issue in Saeed, was whether the police could conduct a penile swab on a man arrested on suspicion of sexual assault (or effectively force him to do it for them) in order to obtain the victim’s DNA ― without a warrant. In an opinion by Justice Moldaver, a seven-judge majority answered that question in the affirmative. Justice Karakatsanis disagreed, but would have admitted the evidence under section 24(2) of the Charter. Justice Abella dissented, endorsing Justice Karakatsanis’ approach to the main issue, but being of the view that the evidence was not admissible.

The friend of a victim of a brutal sexual assault having pointed to Mr. Saeed as the perpetrator, the police arrested him. Some time later, having concluded that a penile swab would potentially uncover useful evidence, the police placed Mr. Saeed, fully clothed, “in a dry cell, with no toilet or running water, to preserve the evidence. Mr. Saeed was handcuffed to the wall to prevent him from licking his hands or otherwise washing away evidence,” [18] for 30 or 40 minutes. Eventually, Mr. Saeed took the swab, under the direction of two male police officers, who were the only persons present. “The procedure took at most two minutes.” [25]

* * *

Justice Moldaver begins his reasons by asserting ― without much of an explanation ―  that “perhaps more than any other search power,” the “ancient and venerable power” of search incident to arrest “is used by the police on a daily basis to detect, prevent, and solve crimes.” [1] This power extends, he concludes, to the taking of a penile swab, although he “agree[s] that the common law power of search incident to arrest must be tailored to protect the enhanced privacy interests involved.” [42]

Despite the privacy implications, the taking of a penile swab is not, in Justice Moldaver’s view, analogous to that of a bodily sample ― which cannot be carried out as a search incident to arrest. The swab does not reveal information about the arrested person’s body ― it aim is to find the DNA of the complainant. It is also not particularly invasive ― or at any rate less invasive than the taking of dental impressions. Finally, the material it serves to collect can be removed quite quickly, whether deliberately or accidentally.

Justice Moldaver emphasizes the usefulness of collecting this material for prosecuting sexual assaults ― or indeed for exonerating an innocent suspect.  “This type of evidence,” he points out, “is highly reliable. It can be crucial in the case of complainants who are unable to testify, such as children, adults with disabilities, or those who have died or suffered serious injuries as a result of the offence or otherwise.” [59] For this reason, to require consent for a penile swab ― as the United Kingdom does ―  “effectively disregards the interests of victims of sexual assault … and all but ignores the public interest in bringing sexual offenders to justice.” [61] As for requiring a warrant, obtaining one takes time ― hours perhaps ― and thus involves “leaving accused persons to wait for an indefinite period in an uncomfortable and potentially degrading position,” “handcuffed without access to water or toilet facilities … in order to preserve the evidence.” [65]

Justice Moldaver cautions that a penile “swab must be truly incident to the arrest, in the sense that the swab must be related to the reasons for the arrest, and it must be performed for a valid purpose” [74]; that there must be reasonable grounds for conducting one; and that it must be conducted in a reasonable manner, to which end he supplies a list of 10 “factors” or guidelines, admonishing the police to proceed expeditiously, to explain the procedure to the arrested person, to respect his privacy to the extent possible, and to keep records. In Mr. Saeed’s case, the police acted consistently with these guidelines, and the evidence they collected is, accordingly, admissible.

* * *

As she had done in Fearon, Justice Karakatsanis takes a much narrower view of the power of search incident to arrest. She is much more concerned about the privacy interests of the accused, and more skeptical of the ability of courts to prevent abuses by supplying guidelines for the police.

For Justice Karakatsanis, a genital swab (notice, by the way, the gender-neutral terminology she uses, in contrast to Justice Moldaver) are no different from “mouth swabs, dental impressions and hair samples [which] cannot be taken as part of searches incident to arrest because they represent too great an infringement of bodily integrity and affront to privacy and dignity.” [99] Indeed,  “[a] swab of the genital area is far more damaging to personal dignity and privacy than a swab of the inside of the mouth or a pluck of hair from the head,” [101] and this is especially the case for a woman. That a genital swab doesn’t serve to collect information about the individual on whom it is conducted does not matter. The affront to the person’s dignity is the key consideration. However, whatever its purpose, “an effect of the seizure is to put the individual’s DNA in the hands of the state.” [104]

Justice Karakatsanis also notes that “if there is no lawful means by which the police could collect the evidence, ever,” ― and there may not be such means to collect the evidence yielded by a genital swab, as it is not clear that a warrant to collect it can be lawfully issued under the Criminal Code

it would not matter how long the evidence lasts.  Nothing would be lost when the evidence disappeared — no state interests would be compromised —because even if the evidence had survived, the police would have had no lawful authority to collect it. [108]

Further, Justice Karakatsanis argues that although it is not clear that it is actually necessary to handcuff a person in a dry cell in order to preserve the evidence while waiting for a genital swab, if it is,

this necessity could not be used to justify the greater affront to dignity that a genital swab would represent.  One indignity cannot justify another.  It would be ironic indeed if [section 8 of the Charter] did not protect individuals from the indignity of genital swabs precisely because it protects them from the indignity of detention in dry cells. [113]

Finally, Justice Karakatsanis is unconvinced that judicially developed safeguards can effectively protect the privacy interests of all those who may come into contact with the police ― and not only the subset of suspects who will be charged and thus have an opportunity to seek to exclude evidence against them. If Parliament wants to authorize genital swabs by statute, it can do so, but the common law power of search incident to arrest does not extend so far.

Justice Karakatsanis ultimately agrees that the evidence against Mr. Saeed should be admitted, because its admission does not bring the administration of justice into disrepute, not least because “the law on this issue was unsettled at the time of this seizure and the police acted on their understanding of the law.” [129] Justice Abella, who agrees with her section 8 analysis, does not agree with this and dissents.

* * *

For my part, I’m inclined to agree with Justice Karakatsanis. She is right that the distinction which Justice Moldaver draws between the swab at issue here and the taking of other bodily samples ― that the penile swab does not yield, or rather is not intended to yield, the DNA of the person on whom it is performed rather misses the point of the prohibition on taking bodily samples. I also think that she is right to focus on preventing unconstitutional infringements of privacy, and right that this is best accomplished by having clear prospective rules, and not lengthy checklist to be applied, if at all, by judges after the fact. Beyond these specific points, I am concerned by the expansion of the power of search incident to arrest ― including to cases where, as here, those searches take place many hours after the arrest, in the secure confines of a police station rather than in the unpredictable environment in the field. It hardly needs to be said that Justice Moldaver’s paean to that “venerable” power does nothing to soothe my worries.

I will end with a couple of thoughts about judging. It is sometimes suggested, in the heat of controversies about the judicial system’s handling of cases of sexual assault, that male judges systematically fail to empathize with the victims, leading to perpetrators getting off the hook. There is no denying that this sometimes happens. But Saeed shows that one should be careful with generalizations. Here, the Court’s five men sign on to an opinion overtly driven, in significant measure, by concerns about the difficulty of prosecuting sexual assaults. Two of the women members of the Court, by contrast, dissent from their approach, in the name of respect for privacy rights.

No doubt, a judge is influenced in part by his or her background and personal experiences. But that influence is surely more complex than a reflex that causes women to react in one way and men in another. Nor is background the only thing that influences a judge. Adjudication, even in cases involving sexual assault, should not be seen through the lens of a zero-sum battle of the sexes ― unless, of course, a specific judge gives us cause to do so in a specific case. Unless the evidence leads us to the opposite conclusion, we should treat judges as thinking human beings ― apt err sometimes, perhaps often ― but thinking all the same, and not mere automatons.

How to Get It Right on Wrongs

Ontario’s Superior Court has created a new tort. But should it have, in the circumstances?

In Doe 464533, 2016 ONSC 541, a delivered a couple of weeks ago, Ontario’s Superior Court of Justice awarded substantial damages to a person whose ex-boyfriend posted an intimate video of her online, in addition to showing it to some mutual acquaintances. He had previously coaxed her into making the video, promising that he would be the only one to see it. The promise was a lie, since he shared the video as soon as she had sent it to him. Justice Stinson finds the defendant liable for breach of confidence, intentional infliction of emotional distress, and public disclosure of private facts. The latter ground of liability had never before been recognized in Canadian common law, prompting David Fraser, at the Canadian Privacy Law Blog, to write that “[t]his is a huge deal.” It may indeed, as Mr. Fraser implies, be a hugely positive development. Certainly in Doe the Court was able to give the plaintiff redress for a blatant and vicious injustice that the defendant had done to her. Nevertheless, I want to ask whether, in the circumstances, recognizing or creating this new cause of action was the right thing to do. I hasten to add that, in asking the question, I do not mean to suggest that it must be answered in the negative ― only that it is worth thinking about.

One reason why I wonder whether Justice Stinson was right to develop the common law of privacy in this case is that, as David Canton has pointed out over at Slaw,

[t]he defendant did not file a statement of defence, and this decision was based on a motion for default judgment. So while the decision is well reasoned, there was no contrary position presented. 

Courts are normally dependent on hearing arguments from both sides of a case. Indeed, Jeremy Waldron has argued compellingly that the clash of competing arguments is an essential component of what law is, at a conceptual level. Needless to say, if one the parties chooses to forego his or her right to make an argument, as the defendant did here, this should not stop a court from ruling and, where appropriate, making an order against that party. But I wonder whether the absence of one of the parties from the litigation is not a reason, a very good reason even, for a court to be reluctant to make its ruling a precedent-setting one.

Now, this concern should probably be heavily discounted if not ignored if the question on which a precedent may be set is unavoidable on the path to a decision on the facts of the case. However, this is not the case in Doe. As I note above, Justice Stinson found the defendant liable for three different torts arising out of the same set of facts. The first two are familiar, although I take it that they had not previously been applied to similar circumstances. Had Justice Stinson “only” concluded that the defendant had committed a breach of confidence and deliberately inflicted emotional distress on the defendant, he could and surely would have awarded the same damages against him. So there was no pressing need for the judge to address the question of whether public disclosure of private facts should be a distinct ground of liability. Of course, we can imagine circumstances where the other causes of action would not be made out, and the new tort would be the only conceivable ground of liability. It is arguable that Justice Stinson’s thoughtful analysis is a service to future litigants. But again, the value of that analysis is lessened by the fact that it does not rest on a full, contradictory argument by the parties.

And then, quite apart from the unusual circumstances of this case, there is the perennial question of the relative roles of courts and legislatures in developing the law and creating new causes of action. Justice Stinson notes, early on in his reasons, that Parliament has criminalized the “publication of an intimate image without consent,” and that one province, Manitoba, has passed legislation to make it tortious. Should the law of other provinces move in the same direction by way of common law development, or should the courts wait for the legislatures ― which presumably are well aware of the problem ― to act? There are certainly good reasons for the courts not to wait. At least so long as any new causes of action are carefully circumscribed, there is probably something approaching a social consensus in favour of granting the victims of cruelty such as that suffered by the plaintiff in Doe a civil remedy. The issues involved are not ideologically controversial, and do not concern complex policy questions on which courts might lack expertise. Indeed, one could go so far as to say that, seeing how clumsy and pernicious legislative attempts to deal with online miscreants can be, courts would do well to act first, so as to remove the temptation to act that legislators would otherwise feel. All the same, it might have been useful for the court to address the question. Justice Stinson’s reasons do not.

To repeat, all that is not to say that Justice Stinson was wrong to develop the law in the way he did, or that his reasoning on the substantive issue of whether to create a tort of public disclosure of private facts could or should not be endorsed by a court considering the matter in a subsequent case. My point is only that we should take his decision as an occasion for a reflection on the courts’ when faced with novel legal issues.

The Power of Google, Squared

I wrote, I while ago, about “the power of Google” and its role in the discussion surrounding the “right to be forgotten” ― a person’s right to force search engines to remove links to information about that person that is “inadequate, irrelevant or excessive,” whatever these things mean, even if factually true. Last week, the “right to be forgotten” was the subject of an excellent, debate ― nuanced, informative, and with interesting arguments on both sides ― hosted by Intelligence Squared U.S. I encourage you to watch the whole thing, because there is really too much there for a blog post.

I will, however, sketch out what I think was the most persuasive argument deployed by the opponents of the “right to be forgotten” ― with whom, admittedly, I agreed before watching the debate, and still do. I will also say a few words about the alternative solutions they proposed to what they agreed is a real and serious problem ― the danger that the prominence of a story about some stupid mistake or, worse, an unfounded allegation made about a person in search results come to mar his or her life forever, with no second chances possible.

Although the opponents of the “right to be forgotten,” as well as its proponents (I will refer to them as, simply, the opponents and the proponents, for brevity’s sake), made arguments sounding in high principle as well as more practical ones, the one on which the debate mostly focused, and which resonated most with me concerned the institutional arrangements that are needed to implement the “right to be forgotten.” The way it works ― and the only way it can work, according to one of the opponents, Andrew McLaughlin (the CEO of Digg and a former Director of Public Policy for Google) ― is that the person who wants a link to information about him or her removed applies to the search engine, and the search engine decides, following a secretive process and applying criteria of which it alone is aware. If the request is denied, the person who made it can apply to privacy authorities or go to court to reverse the decision. If however, the request is granted, nobody can challenge that decision. Indeed, if the European authorities had their way, nobody would even know that the decision had been made. (Telling the owner of the page to which a link is being delete, as Google has been doing, more or less defeats the purpose of the “right to be forgotten.”)

According to the opponents, this has some very unfortunate consequences. For one thing, the search engines have an incentive to err on the side of granting deletion requests ― at the very least, this avoids them the hassle of fighting appeals. One of the proponents, Chicago professor Eric Posner, suggested that market competition could check this tendency, but the opponents were skeptical that, even if users know that one search engine tends to delete more links than another, this would make any noticeable difference to its bottom line. Mostly, the proponents argued that we can rely on the meaning of the admittedly vague terms “inadequate, irrelevant or excessive” to be worked out over time, so that the decisions to delete a link or not become easier and less controversial. But another consequence of the way in which the “right to be forgotten” is implemented would actually prevent that, the opponents, especially Harvard professor Jonathan Zittrain argued. Since nobody can challenge a decision to delete a link, the courts will have no opportunity to refine the understanding of the concepts involved in the “right to be forgotten.” The upshot is that, according to the opponents anyway, the search engines (which, these days, mostly means Google) end up with a great deal of unchecked discretionary power. This is, of course, ironic, because the proponents of the “right to be forgotten” emphasize concerns about “the power of Google” as one of the reasons to support it, as typically do others who agree with them.

If the opponents are right that the “right to be forgotten” cannot be implemented in a way that is transparent, fair to all the parties concerned, at least reasonably objective, and does not increase instead of the checking “the power of Google,” what are the alternatives? The opponents offered at least three, each of them interesting in its own way. First, Mr. McLaughlin suggested that, instead of a “right to be forgotten,” people should have a right to provide a response, which search engines would have to display among their results. Second, we could have category-specific measures directed at some types of information particularly likely to be prejudicial to people, or of little public interest. (It is worth noting, for example, that in Canada at least, we already do this with criminal court decisions involving minors, which are anonymized; as are family law cases in Québec.) And third, Mr. McLaughlin insisted that, with the increased availability of all sorts of information about everyone, our social mores will need to change. We must become more willing to forgive, and to give people second chances.

This is perhaps optimistic. Then again, so is the proponents’ belief that a corporation can be made to weigh, impartially and conscientiously, considerations of the public interest and the right to “informational self-determination” (which is, apparently, the theoretical foundation of the “right to be forgotten”). And I have argued already that new social norms will in fact emerge as we get more familiar with the internet environment in which we live, and in which our digital shadows are permanently unstuck in time. In any case,what is certain is that these issues are not going to go away anytime soon. It is also clear that this Intelligence Squared debate is an excellent place to start, or to continue, thinking about them. Do watch it if you can.