return to ICG Spaces home    ICG Risk Blog    discussions    newsletters    login    

ICG Risk Blog - [ Semi-autonomous "killer robots" are already within reach of asymmetrical attackers ]

Semi-autonomous "killer robots" are already within reach of asymmetrical attackers


As part of my work revolves about inverting toys, technical gadgets, and industrial "found objects" into asymmetrical weapons, I was attracted to Noel Sharkey's presentation at RUSI's The Ethics of Autonomous Military Systems as well as his earlier efforts in venues such as Robot Wars and Techno Games. I have come to see Sharkey inhabiting the intersection of engineering, the application of engineering and ethics of application:

Most robots currently in combat are extensions of human fighters who control the application of lethal force. When a semi-autonomous MQ-1 Predator self-navigated above a car full of al-Qaida suspects in 2002, the decision to vaporise them with Hellfire missiles was made by pilots 7,000 miles away. Predators and the more deadly Reaper robot attack planes have flown many missions since then with inevitable civilian deaths, yet working with remote-controlled or semi-autonomous machines carries only the same ethical responsibilities as a traditional air strike.

But fully autonomous robots that make their own decisions about lethality are high on the US military agenda. The US National Research Council advises "aggressively exploiting the considerable warfighting benefits offered by autonomous vehicles". They are cheap to manufacture, require less personnel and, according to the navy, perform better in complex missions. One battlefield soldier could start a large-scale robot attack in the air and on the ground.

One should never underestimate the lift of a headline grabbing title; A brief Reuters item called, Killer robots pose latest militant threat, have recently ricocheted Sharkey's concerns around the web:

[Sharkey] believed falling costs would soon make robots a realistic option for extremist groups. Several countries and companies are developing the technology for robot weapons, with the U.S. Department of Defense leading the way...

"How long is it going to be before the terrorists get in on the act? With the current prices of robot construction falling dramatically and the availability of ready-made components for the amateur market, it wouldn't require a lot of skill to make autonomous robot weapons." Sharkey said a small GPS-guided drone with autopilot could be made for about 250 pounds ($490).

Writing to Sharkey:

I support your contention and submit that it will happen sooner that the high street press assumes and, if previous al Qaeda operational practices are any guide, robots will come in swarms to both confuse and overwhelm defenders and maximize target damage. [email]

I cited a trio of short weblog items I wrote in April 2004 in pursuit of Commercial-Off-The-Shelf (COTS) fleet of attack and surveillance UAVs (Unmanned Aerial Vehicles):

Price, performance and accessibility have only accelerated since. Subsequent to my articles, the Israeli IDF was astonished when Hezbollah launched a reconnaissance UAV over Israeli territory, recovering it without incident. Despite Israeli drone and UAV flights over Lebanon, Israel had not been paying attention to asymmetrical UAV development - publicly stated in many cases to rise from commercial radio-controlled (R/C) model aircraft versions. They should not have been surprised; Hezbollah is a resourceful adversary.

Constructing, in some instances assembling, a semi-autonomous "killer robot" is all too easy. Remember this effort to construct a COTS fleet of attack and surveillance UAVs was early 2004:

I am not an R/C pilot so I could start clean as would any other reasonably technically inclined individual. My ground rules were:

  • Could pay cash for everything
  • Could buy everything in-country and so not have to bring items across a border
  • Could buy all items in a population-dense environment not immediately likely to be surveilled
  • Could obtain PC-based simulators in order to covertly learn how to pilot either fixed wing or rotary wing aircraft, i.e., before I tried to fly a physical device
  • All essential components were either genuinely plug and play or already available in kitted form
  • Could obtain functional schematics and instructions for all installs/add-ons
  • Ability to install GPS autopilots with ground pilot override
  • Ability to install real-time video cameras and their RF links
  • Ability to install digital camera triggering
  • Ability to carry payloads (and either release, spray, or otherwise distribute the payload)
  • Option for stealth/noise abatement
  • Ability to do it at modest cost in comparison to anything a military unit would field and, labor costs aside, be within al Qaeda's frugal pocket book

I found that as early as 2004, "it is feasible for a diligent and reasonably agile individual or small group to create a COTS hunter-killer and surveillance R/C model fleet, a poor man's Predator":

Ability to assemble an R/C craft that could launch conventionally, switch over to GPS autopilot, fly a course either to a target or a race track round trip and allow it to again be taken over by another user for terminal homing or landing... Many PC simulators [are available] for a variety of fixed wing and rotary wing R/C models.

Nose video cameras that could superimpose imagery over a heads-up cockpit display based on telemetry sent back from the bird. If the ground pilot was properly trained, it was possible to fly something onto the target just like the big boys...

Smoke systems intended for demonstration flying are intriguing as a dispersal mechanism for other agents. Certain smoke pumps use one TX-RX channel to toggle on/off...

If the intent is to surveil or deliver/spray a payload, then an R/C aircraft can be launched, perform its mission, and subsequently be recovered -- if for no other reason than to forestall discovery of the means of an attack or that an attack had occurred. The cost of the systems is low enough and simple enough that it could be produced in a quantity that would satisfy the redundancy needs of groups like al Qaeda.

These small UAVs can have enormous consequences beyond delivery of conventional explosives. Our research into the feasibility of producing asymmetrical small volume, "off scope" organophosphates (nerve agents), i.e., agent production using easily purchased materials and not the more rarified "Australia Group" components, showed that production was not limited to sovereign state actors. See:

Some of our findings: If you are going to make and use an organophosphate product in less than a year, standard stainless steel components will suffice before corrosion degrades the system, inadvertently venting product. Toxic byproducts of production can be exhausted directly into a sealed running water stream, sending it off for the sewer system to absorb. Use of microreactors and microfactory components vastly lower production risks while improving weaponization and delivery.

An article is forthcoming on criteria for an asymmetrical air force that would be within the means of a number of entities, criminal and terrorist.

'Robot arms race' underway, expert warns
Tom Simonite news service
12:10 27 February 2008

The Ethics of Autonomous Military Systems
Royal United Services Institute for Defence and Security Studies
27 February, 2008

Killer robots pose latest militant threat-expert
Tue Feb 26, 2008 7:00pm EST

Robot wars are a reality
Armies want to give the power of life and death to machines without reason or conscience
Noel Sharkey
The Guardian
August 18, 2007

Hezbollah sends drone over Israel
AFP/ABC (Australian Broadcasting Corporation)
Last Update: Monday, November 8, 2004. 9:50pm (AEDT)

Gordon Housworth

InfoT Public  Risk Containment and Pricing Public  Strategic Risk Public  Terrorism Public  Weapons & Technology Public  


  discuss this article

<<  |  July 2020  |  >>
view our rss feed