Here is a post by TTAC21 member Maziar Homayounnejad, who is studying for a PhD at Kings College London. His thesis if focussed on the legal issues of autonomous weapon systems.
This post looks at how breakthroughs in quantum physics could be applied to military technologies, specifically how it could lead to more accurate and discriminate targeting. Quantum physics (explained here) is basically a way of looking at particles in two states at once, for example seeing sun-light as waves, or photons.
How quantum physics applies to computing was rather wonderfully explained by Canadian Prime Minister Justin Trudeau at a recent press conference:
Anyway, here’s Maziar:
For the first post, I thought I’d look at some of the security and defence applications of quantum physics (QP), with specific reference to the ‘Technology Quarterly’ section of this week’s Economist. The supplement contains a series of articles on the coming of age of QP, and it gives us some clues on the direction that military innovation will soon be taking.
After decades of being little more than an interesting set of theories, practical applications of QP are coming about and will continue to advance rapidly in the near-term. This is partly because of advances in various bits of hardware, which are needed for quantum capabilities to be put into effect.
Specific to counter-terrorism and defence, the article ‘Metrology – Sensing Sensibility’ is particularly interesting. It discusses how QP will boost the capabilities of all sorts of sensory devices, such as atom interferometers (a kind of gravity sensor). Five applications are immediately apparent, some of which are in the article. Firstly, accurate gravity sensing will enable attacking forces to detect underground and undersea movement. This could mean the deterrent effect of the UK’s submarines and torpedoes could disappear, unless of course a counter-measure of some sort is discovered.
Secondly, in urban battlefields such as those in Gaza, insurgents and their military supplies often travel through underground tunnels and they maintain underground shelters. Here, gravity sensors may provide valuable intelligence to both air and ground attack units, enabling them to intercept insurgents at the precise point of exit, or to attack en route to pre-empt both the insurgents and the underground movement of supplies. Of course, a measure of precaution is needed as some tunnels may be accessible to civilians, rendering the battlefield status of underground movements less certain.
Thirdly, quantum gravimeters can precisely map geological features from the gravitational force they induce, thereby enabling military units to navigate themselves in areas where satellite signals are weak (or in ‘GPS-denied’ environments). The article cites a British MoD scientist who aptly refers to this as “a kind of Google maps for gravitation”. This will be enormously important for the viability of lethal autonomous weapon systems (LAWS), which may at times have to operate in denied environments, or may have to shut off their own communication links to avoid enemy hacking.
Fourthly, the article goes on to explain that, as gravity is an example of acceleration, gravity sensors are effectively acceleration sensors too. This is a boon to companies developing driverless vehicles, for which accurately sensing movements in their external environment and effective collision-avoidance will be a ‘make or break’. The same applies to LAWS, not just for collision avoidance – important as that is in a potentially chaotic battlefield – but also to make crucial assessments on the status of enemy forces. In some situations, the speed and acceleration of enemy movement towards you is one indicator (and often a compelling one) in determining the likelihood of ‘hostile intent’. The faster the enemy is approaching and accelerating towards you, the more likely you will be attacked; the slower the movement and, especially, if that movement is decelerating, the less likely an attack may take place; retreat, or movement away from a LAWS unit, may even be taken as a sign of surrender, possibly requiring an autonomous system to be programmed to hold fire.
Finally, the article mentions some of the civilian construction applications of gravity sensors, where contractors currently dig holes in roads and other plots of land but without really knowing what’s underneath; pebbles, pipes and underground wiring can all look the same to today’s rudimentary technologies. Consequently, ‘underground surprises’ cause around half of all construction and roadwork delays, though this can be completely avoided with accurate gravity sensors based on QP. In the military sphere, this has obvious applications for collateral damage estimation (CDE). Currently, data-intensive CDE methodology is both sophisticated and accurate, but only in relation to what it can detect. Namely, by sensing the number and size of buildings, the likely material components of those buildings, and other objects and explosives around an intended strike site, and by applying this to the blast radius of the intended munition, CDE methodology can determine both a collateral effects radius and the severity of damage within that radius. From there, the system detects and includes the number of people within the radius to produce a relatively accurate final assessment of collateral damage (including incidental injury to civilians). However, in cases where chemicals, explosives and other collateral effects-inducing objects are concealed or otherwise not taken into account, the CDE methodology will underestimate the true level of collateral damage that actually occurs from an attack. Gravity sensors will go some way to addressing this, enabling those who are planning an attack to take into account underground objects, which may include heavy metallic items or even pipes liable to release dangerous forces, such as gas. It should be noted that the current situation of imperfect knowledge does not affect the legal assessment of proportionality, which hinges on the ‘expected’ and not the actual collateral damage caused. Nonetheless, being better-informed about underground objects that may increase the CDE assessments will undoubtedly improve the decision-making of commanders and give them the option of avoiding PR disasters from heavy civilian casualties.
Aside from gravity-sensing, QP has a whole host of other military applications, from communications and encryption to guard against cyber-attacks; to quantum computing and its effects on machine learning (the difference between quantum computers and today’s supercomputers apparently being like the difference between human intelligence and that of a jellyfish!). Both of these are treated in separate articles in the Technology Quarterly. One more fascinating application of QP that’s worth briefly mentioning is ‘ghost imaging’. Put simply, this combines pictures of a target object (along with all the heat- and smoke-based distortions generated by the fog of war) with light beams reflected directly from the target. Correlating those measurements derives an artificially-generated, but vastly improved image of an object that might be two or so miles away on a smoky battlefield. Essentially, the system is computing the paths that light takes to the target and back to the sensors, and it corrects for distortions on the actual image. Undoubtedly, this will allow machines to more accurately classify persons as either combatants or civilians (as well as objects as being of a civilian or military nature) on a battlefield, thereby enabling it to fulfil the distinction task more accurately (all else being equal).
One thing that’s difficult to deny from all of this is that technological improvements are racing ahead in ways that can be unpredictable. Specifically in relation to LAWS, some of these capabilities can be integrated into autonomous systems, which may or may not be able to take lethal action more discriminately and proportionately than humans on the battlefield. This all points towards a strong preference for regulating technological standards; in particular, for lawyers to set minimum or baseline requirements that a LAWS will have to meet before it is allowed to operate autonomously and without contemporaneous human control. The current approach taken by some non-governmental organisations (i.e. advocating for a comprehensive and pre-emptive ban on LAWS) is ill-informed and often based on pre-emptively negative judgments about future technological capabilities, which of course they are not qualified to do. Not only is this a haphazard way to approach the situation, but to end the debate there could also possibly sacrifice the opportunity to bring greater precision onto the battlefield, in the same way that the advent and further refinement of precision guided homing munitions has massively reduced collateral damage over the past 70 years or so.