Western military theorists and researchers are increasingly preoccupied with how  the rapid growth of cities in the global south undermine their technological advantages over non-state insurgents. In particular, a concerted effort is being made to redesign and reequip the United State military so that its raison d’ êtré becomes the violent take over and control of the mega cities of the global south. After three decades when the US military concentrated on global surveillance, power projection, and the avoidance of ground fighting in urban environments,  US high-tech surveillance and targeting power is now being redesigned to actually target and control the very micro geographies of global south cities.

With the bloody results of the urban insurgency in Iraq adding evidence to support their views every day, many leading military theorists in the US now argue that the urban terrain in poor, global south countries is a great leveller between high-tech US forces and their low-tech and usually informally organised and poorly equipped adversaries. The complex and congested terrain below, within, and above cities are thus widely viewed by US military commanders and theorists as a set of spaces which limit the effectiveness of the high-tech space-targeted bombs, surveillance systems, and automated, ‘network-centric’ and ‘precision’ weapons developed during the Cold War. A US Defense Intelligence Reference Document for example, argues  that ‘the urban environment negates the abilities of present US military communications equipment.’  This results in dead spots which severely undermine the principles and technologies of ‘network-centric warfare” the style of high-tech targeting  and killing that is the preferred mode of operation amongst the US military these days.  Global south cities are thus seen to be refuges which shelter insurgent groups from the overwhelming technological superiority of US forces. The major military think tank RAND report ed recently that this is leading to what they call the ‘urbanization of insurgency’.’

The result of such debates within the US military are major programmes designed to develop high-tech weapons systems which are deliberately designed to be most effective in global south cities and so remove the function of such places as refuges from the high-tech killing power of US forces. Using the usual euphemisms and techo-babble of the military, Major Lee Grubbs  of the US Army argued recently in a US military report that U.S. forces need to be redefined so that their main purpose is, in his words, to:
‘create operational shock in the urban environment. Operational shock as an urban operational concept depends on selective influence. The utility to selectively influence depends on a deep understanding of the battlespace to identify causality between critical point, action, and effect achieved. The level of situational understanding within the infinite relationships of any enemy system and the urban area requires a variation on the doctrinal development of an understanding of the city and operational design. Operational design and a process for understanding the city becomes critical for the selection of critical points to destroy, control and influence. The challenge is the development of an executable operational concept for achieving systematic, across the entire system, effects within the urban environment through the selective use of force’

Two major areas of work have resulted from such a policy shift. The first involves programmes designed to saturate global south cities with small scale surveillance systems so that  US forces can build up a full picture of global south cities in real time. The dream of US military theorists is that this can be done to such an extent that any identified target can be automatically identified at any time and so exposed to high-technology tracking and killing powers of their ‘network-centric’ weapons. Such visions imagine pervasive and interlinked arrays of ‘loitering’ and ’embedded’ sensors overcoming all the limits and interruptions that mega-city environments place in the way of successfully implementing networks centric warfare.

‘Combat Zones That See’

One major example of such a development is the tellingly title ‘Combat Zones That See’ project. This is led by the  US Defense Advanced Research Projects Agency (DARPA) ‘ the agency responsible for dreaming up futuristic military systems.  Launched at the start of the Iraq insurgency in 2003, the launch report for the project claims that it ‘explores concepts, develops [computer] algorithms, and delivers systems for utilising large numbers (1000s) of video cameras to provide the close-in sensing demanded for military operations in urban terrain’. Through installing thousands of computerised videos cameras across whole occupied cities, the project organisers envisage that CTS will enable  ‘motion-pattern analysis across whole city scales.’In other words, the movements of cars across whole occupied cities will be tracked in real time.

A direct response to the ways in which global south cities interrupt the signals necessary for earlier generations of high-tech weaponry, it is envisaged that, once it has been developed by 2007, CTS  ‘will generate, for the first time, the reconnaissance, surveillance and targeting information needed to provide close-in, continuous, always-on support for military operations in urban terrain.’It will be designed to specifically address the  ‘inherently three-dimensional nature of urban centres, with large buildings, extensive underground [passageways, and concealment from above’.

The central challenge of CTS, according to DARPA, will be to build up fully representative data profiles on the ‘normal’ movement patterns of entire subject cities so that computer software could then use statistical modelling to  ‘determine what is normal and what is not’. This will be a purported aid to identifying insurgents’ activities and real or potential attacks, as well as warning of the presence or movement of target or suspect vehicles or individuals.

After a stream of protests from US civil liberties groups, DARPA stressed that, whilst the initial test of mass, urban tracking will take place at a US Army base within the United States (Fort Belvoir, Virginia), the deployment of CTS will only take place in what DARPA call ‘Foreign urban battlefields’.

Saturating occupied or target cities with tiny sensors and cameras is also being investigated an associated DARPA programme labelled HURT. This programme centres on the development of a wide range of unmanned surveillance and weapons platforms tailored to loiter for long periods within and above global south urban environments (Figure). DARPA’s HURT and CTS programmes are, in turn, being backed up by major virtual simulations of widescale future urban wars in cities like Jakarta (an exercise known as ‘Urban Resolve’). In these, future suites of surveillance systems, like those under development in HURT, are inputted into the simulations to assess their likely effectiveness.

 
Figure DARPA’s urban ‘Reconnaissance, Surveillance and Target Acquisition’ (RSTA) platforms as envisaged by its HURT Programme (Darpa, 2004).(LOS=Line of Sight, RSTA= Reconnaissance, Surveillance and Target Assessment), UAV= Unmanned Aerial Vehicle)

Fantasies of Robotised Killing

The second area of work involves a shift towards robotic air and ground weapons which, when linked to the ‘persistent’ surveillance and target identification systems just discussed, will be deployed to continually and automatically destroy purported ‘targets’ in potentially endless streams of state killing. Here, crucially, fantasies of military omniscience and omnipotence, which blur seamlessly into wider sci-fi and cyberpunk imaginations of future weapons, become indistinguishable from major US military Research and Development programmes. The fantasies of linking all-seeing surveillance — which bring God-like levels of ‘situational awareness’ to US forces attempting to control global south megacities — to automated machines of killing, are a central feature here.

Excited by the reports on DARPA’s CTS Program, Defense Watch magazine provide a telling example of such fantasies. In their scenario, swarms of tiny networked sensors pervade the ‘target’ global south city. These provide continuous streams of target information to arrays of automated weaponry. Together, these systems produce continuous killing and ‘target’ destruction: a kind of robotised counter-insurgency operation with US commanders and soldiers doing little but overseeing the systems from a safe distance. (safe, that is, for them). 

Defense Watch thus fantasise about ‘a battlefield in the near future’that is wired up with the systems which result from the CTS programme and its followers. The text is so revealing it is worth quoting at length.

‘Several large fans are stationed outside the city limits of an urban target that our [sic] guys need to take’, they begin:
‘Upon appropriate signal, what appears like a dust cloud emanates from each fan. The cloud is blown into town where it quickly dissipates. After a few minutes of processing by laptop-size processors, a squadron of small, disposable aircraft ascends over the city. The little drones dive into selected areas determined by the initial analysis of data transmitted by the fan-propelled swarm. Where they disperse their nano-payloads.’

 ‘After this, the processors get even more busy’, continues the scenario:
‘Within minutes the mobile tactical center have a detailed visual and audio picture of every street and building in the entire city. Every hostile [person] has been identified and located. From this point on, nobody in the city moves without the full and complete knowledge of the mobile tactical center. As blind spots are discovered, they can quickly be covered by additional dispersal of more nano-devices. Unmanned air and ground vehicles can now be vectored directly to selected targets to take them out, one by one. Those enemy combatants clever enough to evade actually being taken out by the unmanned units can then be captured of killed by human elements who are guided directly to their locations, with full and complete knowledge of their individual fortifications and defenses. When the dust settles on competitive bidding for [the ‘Combat Zones That See’ programme], and after the first prototypes are delivered several years from now, our guys are in for a mind-boggling treat at the expense of the bad guys’

Such fantasies extend even further to the automated surveillance, through brain scanning, of people’s inner mental attitudes to any U.S. invasion, so that ‘targets’ deemed to be resistant can be automatically identified and destroyed:
‘Robotic systems push deeper into the urban area. Behind the fighters, military police and intelligence personnel process the inhabitants, electronically reading their attitudes toward the intervention and cataloguing them into a database immediately recoverable by every fire team in the city (even individual weapons might be able to read personal signatures, firing immediately upon cueing. Smart munitions track enemy systems and profiled individuals. Satellites monitor the city for any air defense fires, curing immediate responses from near-space orbiting ‘guns’. Drones track  inhabitants who have been ‘read’ as potentially hostile and ‘tagged’’

 
Emerging Robotised Killing Systems

Disturbingly, such fantasies are far from the realms of sci-fi fantasy. Rather, as with the CTS and HURT programmes, they are fuelling very real multimillion dollar research and weapons development programmes aimed at developing ground and aerial vehicles which not only navigate and move robotically, but which select and destroy targets without ‘humans in the loop’ based on algorithmically-driven ‘decisions’.

Writing in the US military Signal magazine in 2004,  US Air Force commentator Maryann Lawlor, for example, discusses the development of ‘autonomous mechanized combatants’ for the US Air Force. These are being designed, she notes, to use ‘pattern recognition’software for what she calls ‘time-critical targeting’. This means that automatic sensors will link very quickly to automated weapons so that fleeting ‘targets’ both within and outside cities can be continually destroyed. Such doctrine is widely termed ‘compressing the kill chain’ or ‘sensor to shooter warfare’ in US military parlance .  The ‘swarming of unmanned systems’project team at US forces JOINT Command Experimentation Directorate, based in Suffolk, Virginia, he states, are so advanced in such experimentation that ‘autonomous, networked and integrated robots may be the norm rather than the exception by 2025’.

By that date, Lawlor predicts that ‘technologies could be developed that would allow  machines to sense a  report of gunfire in an urban environment to within one meter, triangulating the position of the shooter and return fire within a  fraction of a second’  providing a completely automated weapon system devoid of human involvement. She quotes Gordon Johnson, the ‘Unmanned Effects’ team leader for the US Army’s ‘Project Alpha’, as saying of such a system that:
  ‘if it can get within one meter, it’s killed the person who’s firing. So, essentially, what we’re saying is that anyone who would shoot at our forces would die. Before he can drop that weapon and run, he’s probably already dead. Well now, these cowards in Baghdad would have to play with blood and guts every time they shoot at one of our folks. The costs of poker went up significantly. The enemy, are they going to give up blood and guts to kill machines? I’m guessing not’

Tellingly, Lawlor predicts that such robo-war systems will ‘help save lives by taking humans out of harm’s way’; here, tellingly, only US forces are considered to fall within the category ‘human’. Their ‘targets’, presumably, are something else entirely.

In addition, unmanned aerial vehicles armed with ‘intelligent munitions’ are already being designed which will, eventually, be programmed to fire on, and kill, ‘targets’ detected by US Force’s real-time surveillance grids, in a  completely autonomous way. Such munitions will loiter over targets for days at a time, linked into the data links, until ‘targets’ are detected for destruction. A programme called TUDLS ‘ or ‘Total Urban Dominance Layered System ‘ for example, is currently underway to provide what Benjamin Plenge  of the US Air Force’s Munitions Directorate in Florida, describes in his geekspeak as:
‘long hover and loiter propulsion systems, multidiscriminant sensors and seekers, mini- and micro-air vehicles, mini-lethal and non-lethal warheads, autonomous and man-in-the loop control algorithms, and a strong interface with the  battlespace in formation network’

Plenge stresses further that the ‘loitering’ munitions developed through the TUDLS programme  will ‘be capable of completing the entire kill chain with minimal human involvement’. They will be able to co-operate to maximise their autonomous destructive power or, where there are ‘more stringent rules of engagement’, through referring back each time they strike to human-in-the-loop ways on working when they are ‘in close proximity to friendly forces’.

Crucially, such munitions will be equipped with algorithms designed to separate ‘targets’ from ‘non-targets’automatically. The ultimate goals, according to Chuck Pinney, an engineer at the Raytheon Corporation, is a ‘kill chain solution’based on ‘1st look, 1st feed, 1st kill’  where each armed unmanned vehicle continuously ‘seeks out targets on its own’. Tirpak, a US air force specialist, envisages that humans will be required to make the decisions to launch weapons at targets only ‘until Unmanned Combat Air Vehicles establish a track record of reliability in finding the right targets and employing weapons properly’.  Then, he predicts, the ‘machines will be trusted to do even that’. 

The bombs that will be tasked with such automated killing are already under development. One, termed LOCAAS (for ‘Low Cost Autonomous Attack Systems’), the USAF has already committed to buy. This loiters and searches over an area of 80 square miles, scanning the area and comparing signals received with what the USAF engineer Greg Marzolf describes as  ‘stored target templates’which use the ‘advanced [computer] algorithms’of what is known as ‘automated target recognition’or ATR software. When the signature of a known ‘target’ is detected, the missile homes in to destroy it.

In Conclusion

Such dreams of continuous, automated amnipotence, of course, must be treated with caution. The US military, and its associated complex of R and D outfits, have, after all, long held fantasies of superweapons which would realise their dreams of mastery and omnipotence . There are many people even within the US military are deeply sceptical that the horrors and ‘fog of war’ in bloody ‘urban operations’ like the Iraqi insurgency, can ever really be automated. With over 1500 US military dead, mostly through bloody and close urban combat, the technophiliac fantasies of US defence contractors and R and D organisations must look highly unrealistic to the US infantry on the streets of Fallujah or Baghdad. As ever,  then, the casualties from the glitzy fantasies of military theorists, who grow ever-fatter on the back of lavish spending increases from the Bush administration, are all of the poor and impoverished of the world’s cities: the US recruits from the ghetto who become the ‘grunts’ who die on the streets of Fallujah, or those invisible and unrecognised subjects who’s bodies pile up, or are maimed, as ‘collateral damage’ on the streets of global south cities, as the increasingly automated killing machines move into ‘action’.

Whether such systems will ever function as imagined even in military terms is, then, beside the point. For, as the death of 100,000 Iraqi civilians within the first 21 months of the US-UK invasion testifies, the very  existence of an imperial project of launching the world’s military hegemon’s high-tech killing systems into global south cities will inevitably lead to mass civilian carnage.  This seems especially so as new computer systems seem likely to emerge which are the actual agents of continuous, autonomous killing as ‘kill chains’ are ‘compressed’, ‘sensors’ are linked automatically to ‘shooters,’ and the fantasies of ‘persistent area dominance’ achieve full expression through the favourable context of Bush’s huge defense spending increases and ideologies of preemptive war.  The real worry is that, in bringing the fantasy that US personnel can be withdrawn from risk to supervise automated machines who do the business of killing in global south cities, the barriers to US aggression might be seen to be further reduced.  Moreover, in suggesting that computer software ca do a better job that an infantryman in separating supposed insurgents from the rest of the population of such cities, the danger is that these fantasies suggest that the ‘collateral damage’ ‘ dead men, women and children civilians ‘ might reduce in such operations.

The broader tragedy, of course, is that these military debates and fantasies translate the human richness of whole global south cities and their residents to mere ‘targets’ to be assaulted and annihilated at will. They reduce the politics of imperial empire to an age-old quest for using the technological advantage of the colonising power  to exterminate those who might politically oppose it. And they fundamentally rest on the essentially racist idea that life in the global south is essentially worthless and expendable.

Email s.d.n.graham@dur.ac.uk

Stephen Graham is  a Professor of Human Geography at the University of Durham. His new book, Cities, War and Terrorism, is out through Blackwell.

---