By Craig Forcese

Full Professor
Faculty of Law
(Common Law Section)
University of Ottawa

Twitter: @cforcese

Subscribe to National Security Law Blog
National Security Law Blog Search
Most Recent Blog Postings

Please also see for Bill C-51-related analyses by Craig Forcese and Kent Roach.


Killing Citizens: Core Legal Dilemmas in Targeted Killing of Cdn Foreign Terrorist Fighters 

My article with Capt. (Ret.) Leah West Sherriff on Canadian and international law and targeted killing is now accepted and forthcoming, Canadian Yearbook of International Law. We have posted the current version here. The abstract is as follows:

For the first time since the introduction of the Canadian Charter of Rights and Freedoms, Canada is an armed conflict with an insurgency that has actively recruited Canadians and directed them to use or promote violence against Canada. In the result, the Canadian government may ask its soldiers to target and kill fellow Canadians, or to assist allies in doing so. This situation raises a host of novel legal issues, including the question of “targeted killing” confronted by the United Kingdom in 2015 when it directed military force against several Britons believed to plotting a terrorist attack. That incident sparked a report from the British Parliament highlighting legal dilemmas. This article does the same for Canada by focusing on the legal implications surrounding a targeted killing by the Canadian government of a Canadian citizen. It examines how a Canadian policy of targeted killing would oblige Canada to make choices on many weighty legal matters. First, it discusses the Canadian public law rules that apply when the Canadian Armed Forces deploy in armed conflicts overseas. It then analyzes the international laws governing military force, scrutinized from the perspective of use of force (jus ad bellum) and the law of armed conflict (jus in bello). It also examines an alternative body of international law: that governing peacetime uses of lethal force. The article concludes by weaving together these areas of law into a single set of legal questions that would necessarily need to be addressed prior to a targeted killing of a Canadian.


Learning Lessons from the Driver Terror Matter

Between them, the Nuttall/Korody case and the Driver take-down raise tremendous questions about Canadian anti-terror strategies. And there is a tremendous temptation to rush to judgment, which can be perilous. On the other hand, there should be judgment and it should be expeditious.

A. Are we a learning society?

Canada is bad at drawing public lessons from past incidents. Lessons, when articulated and even when heeded (sometimes), usually follow time-consuming, expensive, protracted and often delayed (and in the Air India bombing, woefully delayed) commissions of inquiry. Even now, close to two years after the October 2014 attacks, we know very little. The focus of accounting there was on Parliament Hill's security structure, not the broader security and intelligence landscape.

All of this opaqueness can be contrasted with Australia’s detailed seventy-five-page report released shortly after a December 2014 terrorist attack in Sydney.

B. The Risk of Insta-Conclusions

An early judgment from the Driver case might be that we got lucky. The peace bond failed. The tip-off came from the FBI, not a domestic service.  And the intervention from authorities came when the bombing plot was on the cusp of execution. This is too Jack Bauer a plot line for comfort.

Another early judgment might be that authorities got it wrong. The court got it wrong. The people who concluded Driver was not violent got it wrong. The insta-judgment may be that Driver was always dangerous. The conditions should have been tightened, not relaxed. Surveillance should have been unrelenting.

Those are the narratives that risk becoming rooted, which is fine if they are true. But they may not be.

1. Was everyone wrong?

For instance, it could be the case that the authorities were right: Driver was not at high risk of violence when the peace bond was issued. Instead, Driver made the move from radicalized to radicalized-to-violence after the peace bond -- and perhaps (as colleagues Dawson and Amarasingam have suggested) because it unrooted him from the few stabilizing elements in his life, however unpalatable those were. Psychology is not physics. Things change, they are not immutable.

If this is true, it leads to very different policy conclusions than the "everyone was wrong" thesis. Not least, it reaffirms the need to get to grips with counter violent extremism programming and develop best practices in that area.

2. Surveillance on steroids?

There will be very close consideration also of whether the authorities should have maintained more intrusive surveillance. Those will be compelling calls. And more intrusive surveillance could have been facilitated by a more intrusive peace bond.

The conditions imposed, for instance, on Mr. Harkat in an (analogous) security certificate were, at one time: fitting of an electronic monitoring bracelet, to be worn at all times; a strict curfew, with solo departures from his house precluded and accompanied departures only on CBSA approval; strict geographic limits on his ventures outside his house; no visitors other than those listed in the order; no communication with anyone supporting “violent Jihad”; consent to interception by CBSA of all communications; consent to CBSA searches of his house and possessions; surrender of travel documents; and, a bar on possessing weapons.

These conditions seem of the sort that would have forestalled a Driver-type plot.

But there is the difficulty of tailoring conditions to actual risk -- with hindsight certainly, security certificate conditions were overkill, and also deeply controversial.

And more materially (literally), there is cost. By the end of the last decade, the government was budgeting $59 million for two-year periods on security certificates – over $11 million per person on a security certificate. This included more than surveillance costs -- it also included legal costs for the fraught legal contest. But even if you limit these costs to CBSA, CSIS and CIC funding, it still amounts to $3 million per person per year in support of the security certificate.

In February 2016, the government was aware of more than 180 individuals with Canadian connections who were abroad and suspected of engaging in terrorism related activities. One hundred of these were believed to be in Iraq and Syria. Another 60 had returned home. Earlier, in October 2014, the RCMP was reportedly tracking 90 individuals who intended to travel, or had returned from overseas, although it is not clear how many of these were affiliated with Daesh. And these numbers don't include the terror fans who are prepared to be think globally, act locally.

For the sake of argument, let us assume that we impose peace bonds on only 50% of 100 of these potentially dangerous people -- some proportion of the others are prosecuted (even more expensive), die overseas, never return or return and a decision is made that they are not dangerous. (And some of the latter would probably be surveilled anyway, even if no other legal step is taken against them).

Now, given Driver, we decide to apply security certificate-style surveillance to these 50 people. If we assume that the agency costs noted above cover surveillance, that works out to $150 million dollars a year. In 2016-2017, the RCMP's planned budget for all federal major criminal investigations into "serious and organized crime to reduce its impact on Canada and Canadian interests; financial crime to maintain the integrity of the Canadian economy, government and financial systems; and national security threats to support and protect Canada’s national security" is $294 million. So we are talking about adding the equivalent of 1/2 of that total budget in the form of terror-peace bond surveillance costs alone. 

I am not saying this is beyond Canada's means -- the RCMP has a huge budget and CSIS has a big one. And I am not disputing that the $3 million/year cost might be bloated and that policing can be very inefficient. But it is not so bloated and inefficient that there are no real costs at issue. And real costs require real prioritization. 

Keep in mind, meanwhile, that the government has pledged $35 million over five years on counter-violent extremism. This looks like chump change against the costs of persistent surveillance.

So what should our priorities be? I spend more time than I care to admit trying to understand this area of law and policy. But I am not comfortable arriving at a definitive conclusion on the data we have before us.

C. Learning to Learn Lessons Learned

So I circle back to my original point: If this were Australia, we might have a prompt, public lessons-learned report to help us decide and justify, based on facts and not rooted insta-judgment.

This is one practice well worth emulating.





Aaron Driver Matter: Questions awaiting Answers

[Upate: We have prepared two pieces outlining our post-RCMP press conference questions on this matter in the Globe and National Post.]

Yesterday, Aaron Driver was killed during a police anti-terror operation in southern Ontario. RCMP are expected to provide more details today, but early media reports suggest that bombs were involved. See Stewart Bell's Post reporting here. There are now also several overviews of Driver's history (see, e.g., the Globe story here) -- he was well known to police. Most notably, he was subject to an anti-terror peace bond, a mechanism that allows intermediate constraints on liberty on evidence short of what is required to prove an outright terrorism offence. That peace bond was sparked by Driver's online expression of pro-Daesh views and support.

It is a recurring truth that every terrorism incident tends to affirm each person's prior policy preferences. Those who wish tighter laws see them as justifications for tighter laws. (Or in Donald Trump's case, torture). Those who prefer softer forms of anti-terrorism see every incident as a failure to apply properly that approach. Evidence in support of such different positions is often a secondary consideration.

I am vulnerable to that tendency as well, and strive to counter it by self-imposed socratic probing of my own reactions. And so here are some of the questions to which I need answers in order to gauge the law and policy implications of this latest event.

As I write this, I am also cognizant that this is a very human case -- there was loss of life, and by all accounts, there could have been even graver loss of life. Both of those things needs to always be in our minds.

1. Did the peace bond work or did it fail?

For a more detailed account of peace bonds and how they work, see here.

A hasty assumption may be that the peace bond here failed. After all, if the reporting is correct, Driver was well on his way to building a bomb (and may have done so), despite being subject to peace bond limitations.

The peace bond "failed" argument may come in several forms.

a) Prison Uber Alles Theory

First, there will be those who think Driver should have been outright incarcerated a long time ago.  And indeed, when anti-terror peace bonds were created in 2001, they were derided as too weak for real terrorists, and critiqued as potentially too strong for false positives -- that is, innocent persons swept into their embrace by the very low burden of proof on government. Peace bonds are vulnerable to what Kent Roach and I describe as the "Goldilocks" problem.

Incarceration depends, however, on a crime. And our terrorism criminal law already sets the tripwire for terrorism crime very far from actual acts of violence. (None of the two dozen or so persons in prison for post-9/11 terrorism crimes got further than plotting before they were charged and convicted).

So does that mean that we should set the tripwire even further back? In essence, bill C-51 did that by creating a new speech crime -- had it been in force at the time, perhaps it might have reached Driver's speech conduct.

Of course, broadening the sweep of criminal law means that more people will go to jail for ideas that would never, left alone, have resulted in violence -- the terrorist fanclub is many multiples bigger than the terrorist cadre.

There will be those who take no discomfort in the idea of people imprisoned for their hateful thoughts. Personally, I can think of nothing more counterproductive from both a civil liberties and security perspective. The civil liberties implications of imprisoning people for opinions should be obvious. As for security: past practice suggests that putting a radicalized person in prison is a good way of ensuring that person emerge after there sentence radicalized and violent. That seems to be the pattern in recent French terrorist incidents, where imprisonment served as a radicalization-to-violence boot camp.

Canada does not have a sustained prison "deradicalization" program. Nor is it clear that deradicalization is even possible, although terrorist disengagement from violence may be.

So those advocating incarceration as a silver bullet should be careful what they wish for.

b) The peace bond conditions constraining liberty were not stringent enough

This is a more pressing issue. I do not, however, think we know enough to pass judgment. It may have been a mistake to relax the peace bond terms to eliminate (reportedly) an electronic bracelet requirement -- perhaps that was enough to allow Driver a greater range of action. On the other hand, maybe the peace bond conditions provided police with investigative access to Driver greater than would have been the case if reliant on regular surveillance powers. If so, then the peace bond might have facilitated disruption of the apparent plot.

c) The peace bond condition on deprogramming should have been imposed

Driver successfully challenged in court the constitutionality of part of the peace bond law (as amended by C-51) allowing the state to impose counselling -- billed by the court as "deprogramming". Was this decision an example of the constitution getting the way of good security?

Well, the government chose not to appeal the Manitoba lower court decision -- which was rather thin in its reasoning. It is difficult to see this decision from a single lower court as the last word on the matter.

But more than that, there is the question of whether a "treatment program" would have worked. Deradicalization is a novel area of the social sciences, and there is considerable doubt in the literature that I have read as to whether it can ever work. Moreover, forced participation in such a program may have the effect of simply entrenching views.

On the other hand, it may be possible to dissuade people from making the leap to violence. And we do need some sort of exit strategy from an endless cycle of disrupt or incarcerate. A UK parliamentary study (correctly) urged the need for an exit ramp from that country's equivalent to peace bonds, known originally as control orders: "a “graduated scheme, which commences concurrently with the [liberty limiting] measures, with the sole purpose of engagement and de-radicalisation”.

Parking someone on a peace bond is not a permanent solution.

It is possible to see, therefore, the Driver case as underscoring the need for serious investment in counter-violent extremism, so that such matters do not end with an explosion and preventive use of lethal force. Some will scoff at this conclusion, but I do not think the evidence exists one way or another as to whether we can develop an effective counter-violent extremism strategy to preempt these situations. And I don't think we have much choice but to try.

d) C-51 was good! No C-51 was bad!

Already commentators are approaching this case through the optic of the debate on 2015's anti-terrorism law. I am on record with my views about C-51. As careful readers will know, I am not among those who contested C-51's objectives. Rather, as Roach and I discussed in our book (still very bang-on-topic! still reasonably priced!), the law managed to muddle solutions to real problems. We believe it will have negative consequences for civil liberties and security if the powers accorded in C-51 are exercised to their legal limit. (And it is therefore those ill-conceived limits that demand reconsideration).

But in relation to Driver and peace bonds, C-51 so far is a minor cast figure. Peace bonds were not the key controversy in C-51. They have existed since 2001 to respond to feared terrorism offences. The government began its proceedings against Driver under this old law, only switching to the C-51 regime when that law came into force. C-51 did relax the standards for obtaining a peace bond -- they went from very low to very, very low. But I have little doubt that Driver would have been subjected to a peace bond under the original, very low standard -- certainly that was the Crown's belief.

I think that this is more generally true for other peace bonds since C-51 came into force in June 2015. The impediment to peace bonds prior to C-51 wasn't the law -- it was police resources, a different security environment and generally a more conservative approach to using this tool.

As noted, C-51 may have made Driver vulnerable to new speech crime charges, had it been in force. But as also noted, I do not believe incarcerating people who say the wrong things is a good idea.

C-51 also opened the door to CSIS threat reduction activities -- known popularly as "disruption". These include (most controversially) disruption activities that break the law and breach human rights in the Charter. We do not know if any of these new powers were material in this case, although we do know from Minister Goodale's statement that CSIS was involved in this case somehow (probably in its intelligence capacity). To date, there is no reason to believe that CSIS used its law breaking powers. (Nor, incidentally, does this case seem to have any of the controversial hallmarks of the Nuttall/Korody entrapment matter.)

e) CSIS and RCMP Coordinated

There seem to be early indications of police/CSIS coordination, and it is welcome. But the how of the coordination will be interesting, given our longstanding critiques that Canada does not manage inter-agency anti-terrorism investigations as seamlessly as we believe necessary in the current threat environment, or as seamlessly as the UK's MI5 and police do.

The system worked, sort of. And it is important to understand whether this was by design or luck. There will necessarily be questions about whether there was an intelligence/police failure. Here are some bottom-line questions:

  • How did someone known to authorities and subject to a peace bond get as far as posing a credible (and perhaps actual) suicide bomb threat?
  • Why was the police intervention 11th hour, in the nick of time (and perhaps after it, given that there was a reported explosion causing injury)?
  • Given the presence of a weapon and that there was apparently a computer video involved, Driver was in clear violation of his peace bond terms (and so basically automatically subject to imprisonment for up to 4 years). So why was there no arrest much earlier?
  • Why was there no public warning?
  • All of this is to say: Where was the early detection? Was this a resource issue? Did something go missing? Will be discussing a failure to "connect the dots"?

In truth, these are elemental, key questions, and I fear we will not learn the answers. Unlike other countries, we do not do thorough, public lessons-learned reports. We still, for instance, do not know all the details about what went wrong (or right) in the October 2014 terrorism events.

To be clear, I am not advocating gotcha inquiries. Just fact-based, clear-eyed assessments that can be evaluated outside of the hothouse of the security services themselves.

I am told the RCMP will be holding a press conference today at 1:30. In a perfect world, these questions will be asked and answered.

f) Civil liberties groups did their jobs

I add a commentary. I am seeing some suggestions on social media that the civil liberties group that contested Driver's peace bond was wrong to do so. But most of today's rights exist because someone -- often a civil liberties group -- defended them in relation to a deeply unpopular person or group. The core right at issue here was expression, something that enables me to write this blog and tweeps to tweet. 

Defending liberty while preserving security is like balancing an umbrella on the tip of one's finger: it is never in perfect equilibrium. Instead it is always a dynamic exercise in balance involving constant reaction and correction. Without an active espousal of civil liberties values, we risk the umbrella falling forward in constant reaction, and never any correction. Civil liberties groups do their jobs, just as security services do. We need them both to bring their A games.