Before one dismisses this book based on its seemingly pollyanna title, I’d suggest one think of it as an opening volley in what promises to be a series of crucial debates that will play out — one way or another — in the years to come. I believe Smith, founder and CEO of TASER and Axon, did a great job of presenting an argument for the pursuit of a range of technologies and policies intended to curb violence, as well as anticipating, presenting, and debating many of the opposing arguments. The book’s tone is more pragmatic than its bold and controversial title might suggest. That said, I don’t agree with all of the author’s conclusions by any means; though I do agree these questions need to be thoughtfully considered and debated.
I’d put the technologies and policies Smith advocates for into three basic categories. First, those that are nearly inevitable given societal winds of change and the nature of technological development (e.g. nonlethals becoming the primary weapons of the law enforcement community, automated systems being deployed to curb violence in schools, and ending the war on drugs.) Second, those which may be laudable, but which are hard to imagine coming to fruition in the world we live in [or are likely to see in the foreseeable future] (e.g. nonlethals becoming the primary [or exclusive] weapons of the military.) Third, those which are so full of the peril of unintended consequences as to be, frankly, terrifying – if not dystopian (i.e. the use of surveillance and profiling technologies to ACTIVELY attempt to prevent crimes that haven’t yet happened.)
Instead of describing the contents of the book chapter by chapter, I’ll discuss its ideas through the lens presented in the preceding paragraph – starting with the seemingly inevitable technologies. The central thrust of this book is that nonlethal technology needs to be developed / improved such that nonlethals can take up a progressively greater portion of weapons deployment and usage, with the aim of ultimately replacing firearms (and other lethal weapons) with nonlethal weapons. It’s important to note that Smith doesn’t suggest such a replacement could happen at present. He acknowledges that nonlethals are currently not as effective and reliable at incapacitating a threat as are firearms, and he isn’t advocating that people be put at risk by having to defend themselves with an inferior weapon. However, it seems reasonable, given the tremendous technological advances that have occurred, that nonlethal weaponry could become as or more effective than firearms.
If that doesn’t seem reasonable, I would remind one that firearms aren’t – as a rule — as instantaneously and definitively incapacitating as Hollywood portrays. One can find numerous cases of individuals still moving with a magazine’s worth of bullets in — or having passed through — them. (And that’s not to mention the lack of precision that tends to come with throwing a projectile via a controlled explosion.) The point being, one isn’t competing with perfection – so one doesn’t need to be perfect, only better than an existing [flawed] system.
Smith addresses the many dividends of nonlethal weapon usage over that of the lethal counterparts, and there are many. For one thing, killing isn’t easy on anyone (anyone who’s right in the head any way.) Even when a killing is legally justifiable and morally defensible (or even state-sanctioned) it often still results in traumatic stress. For another, there is the reduced cost of getting it wrong, and the adverse societal impacts (e.g. revenge killings) that result from wrongful deaths. Long story short, if one can produce a nonlethal that’s consistently as effective at incapacitating threat, it’s hard to make a rational argument for not fielding said weapon. The example of an automated system to respond to school shootings is an extension of the nonlethal weapons argument, as it’s ultimately based on nonlethals deployed by drone (or robotic system.) The chapter on the war on drugs (ch. 15) bears little discussion as it’s no news that that “war” has been a failure and a phenomenally ineffective way of addressing a societal problem.
That brings us to the laudable but unlikely category in which I put military use of nonlethals as primary (or exclusive) weapons. I’m not saying that military nonlethal weapon systems won’t continue to be developed, improved, and deployed. Given the degree to which war of late features non-state actors and unconventional warfare, it’s possible to imagine such weapons playing a dominant role in specific operations. After all, military members aren’t exempt from the psychological costs of killing. However, military forces deploying into a war zone with nonlethals as their primary weapons is almost impossible to imagine, especially considering the diversity of conditions and opponents for which a military needs to be ready.
In warfare, there is something called the “force multiplier” effect of wounding an enemy over killing an enemy. That is, if you wound someone, it takes two people to carry him or her, plus a chunk of a medic’s time. So, one can imagine four people being out of the fight because one person is severely wounded, versus the one person who would be out of commission (the dead person) if the individual were wounded. To be fair, Smith imagines technology (drones and robots) doing the heavy lifting. Still, it’s hard to imagine how one side in a conflict wins if they have to transport, warehouse, feed, and care for every enemy that is incapacitated while the other side is just killing away. Even if that one side is much more automated, it seems tremendously expensive – even for a relatively small-scale war.
That brings to me chapter five, which I found chilling. That chapter considers how artificial intelligence and surveillance programs (albeit with judicial oversight and other protections) could be used to anticipate crimes so that law enforcement could actively go forth to try to prevent them. (If this sounds a lot like the Tom Cruise movie loosely based on a PKD story, “Minority Report,” it’s because it essentially replaces the three pasty precognitives with computers and offers a bit more oversight. While Smith cautions against taking fictional stories too seriously, he employs some fictional scenarios that I believe might be as a pollyanna as the Spielberg film is dark.) At any rate, the word “actively” is crucial to my concern. I’m all in favor of what has historically been known as “preventive law enforcement” — activities such as putting more patrols in high crime areas, youth mentoring programs, and programs that inform people and businesses about how to be harder targets. However, the idea of police going out and engaging people as though a crime has been committed when none has been conjures images of cities on fire.
First, such an approach is predicated on watching everybody – at least everybody’s online activity – all the time. Which seems both dystopian and of limited effectiveness. [What percentage of people who post on FB that they want to shoot someone are likely to do so?] What about the judicial oversight and related protections? When is a warrant issued to surveil or arrest a person? The warrant is issued based on something an artificial intelligence system already flagged, meaning a government entity is watching everybody’s behavior on a constant fishing expedition. I’m not fond of that idea at all.
Second, we aren’t nearly as good at forecasting the future as we think. Violent crimes are rare and often spontaneous events, and that puts them in classes of behavior we are particularly bad at making predictions about. And, we haven’t eliminated the trade-off between type I and type II error. Imagine there is a question about whether individual X is to be detained based on what the AI spit out. X either was or wasn’t going to commit a crime. We can imagine a four-way matrix in which two of the solutions are correct (i.e. 1.) X was detained and was going to commit a crime; 2.) X wasn’t detained and he wasn’t going to commit a crime.) However, since we can’t know the future [like, at all] the potential remains for mis-estimating whether X was going to commit a crime. So, we have two potential errors (i.e. 1.) X wasn’t detained but he was going to commit a crime [and thus did]; 2.) X was detained but he wasn’t going to commit a crime [wrongful detention].) So, we want to minimize the first error because any violent crime is unacceptable? We go out and shake down more high risk individuals. While we succeed in preventing crimes, we also end up with more wrongful detention. Our legal system’s requirements with regards evidence suggest that as a society we are averse to wrongful disruption of a person’s freedom. Hence, while a “preponderance of evidence” is sufficient for cases where one might lose money in a civil case, if one might be imprisoned, the standard becomes “beyond a reasonable doubt.” Wrongfully detaining an individual when a crime was committed may be sad, but doing it when there is only a suspicion that a crime might likely be committed is tragic.
Of course, under present standards one can’t detain a person for very long. So you let them go, and maybe they do the crime – whether or not they intended to in the first place (ever heard someone say, “if you’re going to treat me like _______, I’m going to act like _______?” I’ll admit that it’s a bit far-fetched but if the system spurs one crime in a million subjects detained that wasn’t going to happen, is that acceptable?) Alternatively, one could place surveillance on the individual. In which case, one is essentially living in Stalin’s Soviet Union. Congratulations. It seems to me this approach offers either huge costs for a marginal gain, or you go full dystopia and knock out crime at a horrifying cost. Neither way seems appealing, but – then again – I am not willing to pay any price to keep anything bad from ever happening to anyone.
I found this book to have some fascinating ideas and to spur my thinking on subjects I might not otherwise have considered. While there was a significant bit that I found unsavory, I also discovered some ideas that were intriguing and worth pursuing. I would highly recommend this book for those interested in issues of technology and policy.