The Real Problem with Lethal Autonomous Weapons Systems (LAWS)

February 9th, 2021

Awake Goy

With the extremely rapid advances in technology made in the twenty-first century, many aspects of human life have transformed irrevocably. One of the most significant changes involves norms regarding the commission of intentional, premeditated homicide by governments. The practice is today termed “targeted killing,” but it differs only in the implement of death from what in centuries past was called “assassination” and deemed illegal. Black-ops involving shady assassins who stalk and eliminate perceived enemies under a cloak of secrecy are no doubt still carried out by governments. But the use of unmanned combat aerial vehicles (UCAV) or lethal drones to stalk and eliminate terrorist suspects in lands far away is openly acknowledged and has been largely accepted by politicians and the populace alike as one of the military’s standard operating procedures.

The use of lethal drones to kill rather than capture suspects began in Israel, but was taken up by the George W. Bush administration in the war on terror waged in response to the attacks of September 11, 2001. President Barack Obama then expanded the practice, electing essentially to eliminate the problem of longterm detention of suspects in facilities such as the prison at Guantánamo Bay by defining them as guilty until proven innocent and then dispatching them using missiles launched from drones. The suspects killed were classified posthumously as Enemy Killed in Action (EKIA) unless specific information demonstrating their innocence was brought to light. But since many of the killings took place in remote parts of the world, such as the Federally Administered Tribal Areas (FATA) of Pakistan, where there were few if any troops or intelligence analysts on the ground to do the sort of due diligence needed to establish the innocence or even the identity of the persons killed, this nearly never happened.

With the ascendance and spread of lethal drones, government officials have effectively permitted the current state of technology to dictate morality, rather than subjecting proposed tools to scrutiny before using them. This is most plausibly a result of the fact that the experts to whom politicians defer on these matters are invariably either military officers or persons with ties to military industry. Indeed, many military officers end up serving on the boards of weapons manufacturing and military logistics firms. The revolving door between government service and industry is evident in cases such as those of Dick Cheney, James Mattis and Lloyd Austin, all of whom served as secretary of defense and also sat on the boards of private military companies with sizable government contracts. From the perspective of military experts, whose focus is upon winning wars through maximizing lethality, the development of remotely piloted aircraft (RPA) has naturally been regarded as a boon, offering the possibility of combating the enemy without risking soldiers’ lives.

Yet in the development and spread of remote-control killing technology, important ethical considerations have been overlooked. First, during regular combat warfare, when troops are placed in dangerous situations, where “kill or be killed” becomes a prudential maxim for survival, many acts of killing can be construed as literal acts of self-defense. Whether or not the troops should be there in the first place, as in Iraq or Vietnam, is another matter altogether, but if a soldier is already in a perilous theater, with enemy combatants lurking around every corner, then the pretext of self-defense becomes reasonable. The same cannot be said for acts of killing perpetrated by soldiers sitting in trailers in Nevada, who are not being directly threatened by their targets.

U.S. combat soldiers on the ground in both Vietnam and Iraq killed many people who might have been insurgents but proved not to be. The veterans of those conflicts suffered enormously as a result, and many ended up permanently wrecked by the experience. Soldiers who use drones to target the enemy are far from the bloody fray and physically safe from the dangers of the “battlefield” on which they fire. Nonetheless, drone and laser sensor operators such as Brandon Bryant abandoned the profession after having become disillusioned with the disparity between what they had signed up to do (defend their country) and what they ended up doing, killing poor tribesmen living out in the middle of nowhere who were not threatening anyone with death at the time when their lives were abruptly ended.

Because drone operators follow and observe their victims for extended periods of time, and witness their anguish in the aftermath of strikes as they bleed out, they have been prone to suffer bouts of regret and develop post traumatic stress disorder (PTSD) despite never having been directly endangered themselves. Such reflective servicepersons furthermore recognize that collateral damage, said to be unavoidable in the “fog of war,” is truly excusable only in a life or death, do or die, dilemma. Up to now, what the drone and laser operators had to fall back on was the fact that they were not in a position to be able to assess the value of the intelligence used to select targets. Their job was to locate and kill the person(s) said to warrant elimination by officers higher up in the chain of command. Accordingly, when mistakes were made, the blame ultimately rested with the analysts who had built the case for targeting on the basis of evidence gathered by drones, obtained through paid informants, and mined from cellphones. In other words, even if the drone operators themselves regretted having killed persons whom they themselves did not believe deserved to die, based on their own observation of the targets, some among them were still able to assuage their conscience by invoking the tried-and-true “invincible ignorance” line, according to which soldiers are not to blame when negative consequences arise from their having executed what to all appearances were legal orders.

But surely intelligence analysts, too, may suffer regret when obviously (or even possibly) innocent people are destroyed on the basis of the analysts’ marshaling and interpretation of the available data. Why not, then, take the fallible human being out of the loop altogether, thus minimizing the possibility of error and the human vulnerability to emotions which sometimes culminates in PTSD? If it was better for soldiers in trailers in Nevada to kill thousands of terrorist suspects throughout the Global War on Terror, rather than having them fly dangerous combat missions, would it not be even better to relieve all parties involved of the burden of having killed?

Despite the moral dubiousness of killing “enemy soldiers” who are not directly threatening anyone with harm, and a fortiori in countries where there are no allied combat soldiers on the ground said to require force protection from above, remote-control killing technology continues to be refined and extended with the aim of making drones both more efficient and more lethal. Consequently, a world in which robots “decide” whom to kill, as in dystopic films of the twentieth century such as Terminator, Robocop and their sequels, is no longer the mere fantasy of writers of speculative fiction. Lethal Autonomous Weapons Systems, with the proper-sounding “LAWS” as its acronym, are currently being pursued as the best way both to keep soldiers off the battlefield and also to minimize the errors invariably committed by all-too-human operators in drone warfare. From a purely tactical perspective, an obvious benefit of LAWS is that with this new technology, which takes human beings “out of the loop,” when mistakes are made, there will be no operator who must bear the burden of knowing that he killed people who did not deserve, much less need, to die. Indeed, arguably the most significant benefit to the military in rolling out LAWS will be the elimination of PTSD among drone operators who deeply regret their participation in the serial, mass killing of persons who posed no direct threat to their killers when they were incinerated by missiles launched from drones.

With LAWS, the responsibility for mistakes made can be almost completely diffused, for computers will not only gather and analyze the data, but also select the targets on the basis of that data, and then launch the missiles themselves. The magnitude of the mistakes made will vary from case to case, but so long as human beings are involved in the construction and programming of the machines used to kill, then the potential for error will obviously remain. There may still be a bit of room left for soul searching among those who programmed the computers, but they will always be able to absolve themselves by pointing to the inherent limitations of data collection. Without perfect information, mistakes will continue to be made, but the lengthier the causal chain, the fewer individuals there will be who feel the need to shoulder any blame.

From a tactical perspective, all of this may sound very logical and clearly better than having soldiers risk their lives, and analysts and operators suffer psychological distress upon learning that they contributed to the carnage when innocent persons are erroneously destroyed. The first premise in the inexorable march toward Lethal Autonomous Weapons Systems, however, that the killing will happen, with or without human operators and analysts, needs to be subjected to scrutiny. What has propelled the mad rush to develop and implement LAWS is the false assumption that the killing ever needed to happen in the first place. The governing idea has been that because the persons being targeted have been determined to be potentially dangerous, they might undertake to threaten people at some future time, if they are permitted to live. In other words, the victims are being preemptively eliminated, following the reasoning used to promote the 2003 invasion of Iraq, when the warmakers claimed that Saddam Hussein posed a threat to the world because of his alleged possession of weapons of mass destruction (WMD). That pretext was of course later found to have been false, along with others, including the claim (obtained through torture) that the Iraqi dictator was somehow in cahoots with al Qaeda. Yet the war went on all the same, with some pundits and war supporters filling the justificatory void with the tried-and-true need to spread democracy.

In the maelstrom of the wars on Afghanistan and Iraq, assassination was simply rebranded as targeted killing, when in fact both practices involve the intentional, premeditated elimination of persons deemed potentially dangerous. This criterion is so vague as to permit the targeting of virtually any able-bodied person who happens to be located in a place where terrorists are suspected to be. The only differences between assassination and targeted killing are the nature of the weapon being used and the fact that soldiers wear uniforms, while undercover assassins and hitmen do not. But are these differences morally relevant?

Unfortunately, over the course of the more than twenty-year Global War on Terror, there has been no attempt to reckon with the facts. But if the war on Iraq was a violation of international law, then every person killed in the conflict was the victim of a crime. Because of the shock of the events of September 11, 2001, however, most of the people who pay for the military’s killing campaigns have gone about their business, allowing the government to use their tax dollars to kill people who had nothing to do with the terrorist attacks, and in many cases were protecting their own land from illegal invaders. Twenty years on, the military continues to kill people when and where it pleases under the pretext of the need to fend off the next terrorist attack. That millions of persons have been killed, maimed, widowed, orphaned, reduced to poverty and/or rendered refugees as a result of the ever-expanding missions of the U.S military in the Middle East and North Africa—most of which were caused by overspill of previous missions, beginning in Afghanistan and Iraq—has been largely ignored.

The “killing machine” has been on autopilot for some time now, in the sense that lists of targets continue to be drawn up and dispatched with the killers themselves writing the history of what transpired. The wars on Afghanistan and Iraq gave rise to new terrorist groups such as ISIS, which then spread to countries such as Pakistan, Yemen, Libya, Syria, Mali, and beyond. Subsequent interventions in those lands then led to the spread of factions throughout Africa, where drone bases have been erected in several countries to deal with the problem of radical Islamist terrorism. With LAWS, the perpetual motion targeting of unnamed persons can be expected to be revved up to run even faster, around the clock, for robotic killers suffer neither compunction nor fatigue, and the success of their missions will continue to be measured by the number of “dead terrorists”, who in fact are suspects. In other words, the ethical problem with LAWS will remain precisely the same as the ethical problem with the drone program through which human operators have pressed the buttons to launch the deadly missiles.

The debate over LAWS should not be over how to make robots act as human beings might. Rather, we must pause and back up to ask why anyone would ever have thought that this rebranding of assassination as the military practice of “targeted killing” should be permitted in the first place. The fallacy in thinking that lethal drones and LAWS “protect the troops” derives from the assumption that the people being killed would have been killed had this technology never been developed. The truth, however, is that the many drone bases now peppering the earth have served as a pretext for launching missile attacks which would otherwise never have occurred. With such tools at their disposal, military and political administrators are apt to use them without thinking through the moral implications of what they are doing, specifically ignoring the long-fought advances in criminal justice made over millennia, above all, the presumption of innocence upheld in free societies the world over.

Drones were originally deployed for surveillance purposes, but it did not take long before they were equipped with missiles to provide a dual-function machine capable of both collecting data and taking out enemy soldiers based on that data. Most of the individuals eliminated have not been identified by name, but in some cases specific persons have been hunted down and killed, as in President Barack Obama’s targeting of U.S. citizens Anwar al-Awlaki and Samir Kahn in Yemen in October 2011, and Prime Minister David Cameron’s killing of British nationals Reyaad Khan and Ruhul Amin in Syria in August 2015. More recently, on January 3, 2020, President Donald Trump targeted top Iranian commander Qasem Soleimani, who was located in Baghdad at the time. Trump openly avowed that the act of killing was intentional and premeditated. According to the president, the major general was responsible for past and future attacks against the United States. All of these eliminations of specific, named individuals would have been considered illegal acts of assassination in centuries past but are today accepted by many as “acts of war” for the simple reason that they are carried out by military drones rather than spies.

The ethical problems with lethal drones have been raised many times by activists, who have protested the killing of persons in countries such as Pakistan, with which the United States is not even at war, and also by successive U. N. Special Rapporteurs on Extrajudicial, Summary or Arbitrary Executions (Philip Alston, Christof Heyns, et al.), who have repeatedly cautioned that the extension of the right to kill anyone anywhere at the caprice of the killers, which has been assumed by the U.S. government in its wide-ranging drone-killing program, can only sabotage the prospects for democracy in lands where leaders opt to eliminate their political rivals, facilely denouncing them as “terrorists” while pointing to the precedent set by the United States, the United Kingdom, and Israel. Needless to say, the literal self-defense pretext does not hold when leaders choose to use remote-control technology to hunt down and assassinate specific persons rather than charging them with crimes and allowing them to be judged by a jury of their peers. But, just as in the case of unnamed targets, when the victims of drone strikes are identified by name, they are invariably labeled terrorists, with no provision of evidence for that claim.

With LAWS comes the specter of fully normalized political assassination with no territorial boundaries whatsoever. The question, then, is not “how do we devise the best algorithms with which to program robotic killers?” Instead, we must ask why homicide should be used in cases where the decision to kill is clearly not a last resort, as it never is in drone killing outside areas of active hostilities, because no human being will perish if the missile is not launched. In expanding the drone program, the Obama administration carried out many “signature strikes”, where the precise identity of the targets was not known but their behavior was said to be typical of known terrorists. In addition, cellphone SIM card data was used to identify persons who had been in contact with other persons already believed to be terrorists or found to have connections to known terrorist groups. To execute persons on the basis of such circumstantial evidence of the possibility of complicity in future terrorist acts is a stunning denial of the human rights of the suspects, and flies in the face of the democratic procedures forged over millennia precisely in order to protect individual persons from being killed at the caprice of those in positions of power. This drone killing procedure in fact exemplifies the very sort of tyranny which finally led Western people to abolish monarchic rule and establish republican constitutions protective of all citizens’ rights. As the mass collection of citizens’ data continues on, such moral concerns are more pressing than ever before, for political leaders may decide to use their trove of intelligence to eliminate not only citizen suspects located abroad, but also in the homeland.

What needs to be done is manifestly not to make machines more efficient and lethal killers. Instead, we need to revisit the first premises which were brushed aside in all of the excitement over the latest and greatest homicide technologies deployed in the twenty-first century, when the U.S. government was given free rein to pursue the perpetrators of the crimes of September 11, 2001. That license to kill with impunity was never revoked, and to this day the drone killing machine continues to be used to destroy people who had nothing whatsoever to do with what happened on that day. With the diffusion of responsibility inherent to LAWS, a truly dystopic future awaits, as the criteria for killing become ever more vague and moral responsibility is fully diffused.