Intelligence of the Future?

Autonomous Systems in the Intelligence Community:

Many Possibilities and Challenges
Jenny R. Holzer, PhD, and Franklin L. Moses, PhD

Introduction • protecting IC systems and net-

works that contain sensitive
The Intelligence Community (IC) information;
has a central mission to help the na-
tion avoid strategic national security • deterring the efforts of others to
Advantages that auton- surprise. Strategic surprise may come acquire US information; and
omous systems could in the form of deliberate actions by
• providing the bases for effective
provide for intelligence adversaries of the United States, or it
and timely security decision- and
may emerge as the result of unantici-
purposes include pated consequences of technological,
policymaking.
movement through economic, demographic, political, or This article discusses “autonomous
varied terrain and en- natural forces. It falls to intelligence technologies” that promise to make
practitioners to find the indicators or humans more proficient in addressing
vironments, stealth, informal signs of change, to organize such needs. Although literature in the
persistent surveillance, knowledge about them, and to identi- field is inconsistent in defining the
and data processing. fy factors influencing their evolution. technology and its components, au-
Practitioners use the information tonomous systems generally are those
to support accurate predictions of that take actions automatically under
future situations and their effects and certain conditions. Put another way,
to provide the basis for appropriate they can be thought of as self-govern-
decisions. ing systems capable of acting on their
own within programmed boundaries.
From its earliest days, the IC has
constantly sought new technolo- An autonomous system may be
gies for intelligence gathering, for platform based—a machine or a de-
counterintelligence activities, and vice such as a robot—or it may reside
for improvements in analyzing and and act entirely in the cyberworld.
interpreting large amounts of diverse Depending on a system’s purposes
data. These new technologies ideally and required actions, autonomy may
would be capable of a number of occur at different scales and degrees
things: of sophistication. In addition, autono-
mous capabilities must be understood
• gathering information from adver-
and developed within the ecology of
saries in inaccessible areas;
specific mission needs, operating en-
• overcoming efforts of adversaries vironments, users, and, in the case of
to deny US and allied access to platform-based systems, the vehicle.1
sources of critical information;
Today’s autonomous systems are in
their infancy, however, capable only

All statements of fact, opinion, or analysis expressed in this article are those of the
authors. Nothing in the article should be construed as asserting or implying US
government endorsement of its factual statements and interpretations.

Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)  21


Intelligence of the Future?

of performing well-defined tasks in tentially toxic chemical facilities, and These characteristics are typically
predictable environments. Advances more. These environments may be adjusted to avoid specific types of
in technologies enabling autonomy dynamic, changing minute to minute. surveillance systems such as human
are needed for these systems to re- While humans may not be able to vision or radar. Autonomy can permit
spond to new situations in complex, navigate safely or undetected to and a system, a surveillance device, for
dynamic environments of the sort through such areas, robots would be example, to quickly and automati-
that most interest the IC. able to do so without endangering cally alter its characteristics when it
individuals, and they could be built detects changes in its environment
to linger in areas of interest for long that might permit an adversary to
periods of time. detect it.
Autonomy’s Poten-
tial in IC Activities Robots could take several forms Persistent surveillance. A high de-
and might include swimmers, gree of autonomy would be required
The key advantages that autono- resembling dolphins, or equipment for long-term surveillance activity,
mous devices could provide for intel- carriers that might look like mules. which would require a surveillance
ligence purposes include movement To be truly effective, however, these vehicle to detect changes in target
through varied terrain and environ- kinds of robots would have to include areas, including spotting and identi-
ments, stealth, persistent surveillance, autonomous systems to automatically fying vehicles in motion. The vehicle
and data processing. control “fins” or “legs” in response to would have to be able to manage
changing water or terrain features. its power and fuel consumption and
Terrain. The types of terrain and communicate with an intelligence
environments encountered on IC mis- Stealth. Systems are made less center and its computers. For exam-
sions vary dramatically. They could detectable by the use of materials ap- ple, an airborne surveillance system
involve desert landscapes, canopied plied to an object’s surface, by physi- might employ an unmanned aerial
forests, crowded urban settings, po- cal properties, and by electronics. vehicle (UAV) with autonomy to
handle flight controls, radar systems,
surveillance equipment, fuel stores,
Early CIA Efforts in Robotic Technology and communications. If a swarm
CIA's Office of Advanced of UAVs were used, then an auton-
Technologies and omous system could control flight
Programs developed
the unmanned patterns and reconfigure the swarm if
underwater vehicle members are lost. A passive bor-
(UUV) fish in the 1990s
to study aquatic robot
der-monitoring system would use au-
technology. Some of tonomy for change detection, motion
the specifications used detection, communications, power
to develop "Charlie"
were: management, and possibly stealth.
-speed
Developed by CIA’s -endurance Data processing. Use of autono-
Office of Research and -maneuverability mous systems for data processing and
Development in the -depth control
1970s, the above micro -navigational distributed computing would present
unmanned aerial accuracy clear advantages for the IC, allowing
-autonomy
vehicle was the first
-communications
for rapid, timely analysis of large
flight of an insect-sized
vehicle (insectothopter). status.3 amounts of data and their incorpo-
It was intended to prove ration into decisionmaking. Auton-
the concept of such omous systems can help humans by
miniaturized platforms
for intelligence doing data analysis at greater speed
collection.2 to fit into a decision cycle. Note well
that autonomy is a resource multipli-

22  Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)

 
Intelligence of the Future?

er, but it does not remove the human may become unmanned under- ments, such as entering and navi-
from the system. water vehicles (UUVs) or UAVs gating through buildings, moving
with autonomous capabilities for underwater, or in environments in
tracking friendly individuals or which there is no access to the global
platforms. positioning system (GPS) and where
In the Hands of Adversaries direct, remote control is not possible.
• Exploiting technological limita- Operational capabilities, howev-
The possibility that enemies of the tions of sensors against them. For er, will require more sophisticated
United States would take advantage example, our sensors may have software than is currently available.
of autonomous systems technology gaps in their electromagnetic (EM) State-of-the-art autonomous capabili-
for their own purposes is high. We spectrums in which data do not ties exist at the subsystem level (e.g.,
must understand what countermea- register. False or malicious code obstacle avoidance for ground robots
sures would be needed to combat may be introduced into the pro- and capabilities to maintain con-
autonomous systems used against the cessing centers to cause systems to trolled flight for air platforms).
United States and its allies. Possibili- fail. An autonomous system’s own
ties to consider include the following: failure modes (i.e., stop transmis- Less well developed is the ability
sions or return to origin) could be to understand and freely navigate
• Surprising US and allied forces used for an enemies’ purposes. an environment. Sensors (hardware)
by using stealth to hide weapons provide the input, but software
or intelligence projects targeted provides the understanding. Simulta-
against friendly autonomous sys- neous location and mapping (SLAM)
tems. For example, sensors can be Enabling Technologies for is an example of a technique that
fooled by changing target objects Autonomous Systems robotic platforms can use to explore
or the environment itself (e.g., their environment and build up a
by increasing turbulence, adding An autonomous system is an inte- three-dimensional (3D) map to deter-
irregular structures, applying dif- gration of enabling technologies that mine their location and navigate.4
ferent paint coatings, using differ- allows it to understand its designed
ent surface materials, or changing goals, sense and understand its However, we must supplement
audible signatures). environment, and make decisions on mapping capabilities with software
actions that it must execute to com- capable of planning courses of action
• Identifying passwords or path- plete its goals. However, a variety of and making decisions to achieve
ways to gain access into sensitive technologies must still be developed them. Such autonomous systems
US systems and networks. Au- (or invented) before the potential of would lessen the burden on IC con-
tonomous systems can be used to autonomous systems is realized. The trollers, who would otherwise have to
combine human intelligence from most easily observed advances in make extensive manual inputs such
Internet social engineering with these enabling technologies are the ro- as GPS way points, sensor tasking, or
distributed processing to identify botic platforms that allow autonomous power management.5
passwords and other obstacles to systems to perform physical actions or
logging into classified systems. move through their environment.
An autonomous system might
take the form of a well-written Researchers are studying biological Evolution of Autonomous Systems
computer virus or malicious code systems to create robots ideally suit-
capable of modifying itself to ed to perform in particular environ- Needed Advances in En-
move undetected through a net- ments. (See examples in the graphic abling Technologies
work and cause system failure. on the next page.) Advances in bio- The technologies that make
mimetics have allowed the creation autonomous systems function must
• Spoofing current sensor networks of new classes of robotic platforms
or creating alternative networks to continue to mature and evolve in or-
designed to perform in complex and der to deal with the complexities and
follow individuals. For example, dynamic assignments and environ-
homemade submarines or aircraft

Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)  23


Intelligence of the Future?

A. The Defense Advanced Research Projects Agency B. A Canadian research group addressed the challenge
(DARPA)-funded BigDog robot uses four animal-like legs
of climbing walls by designing a dry adhesive that
to traverse terrain too rough or slippery for convention-
mimicked the structures of a gecko’s foot pad. The
al vehicles. It is also capable of recovering its “balance” if
platform, called the Tailless Timing Belt Climbing
it slips. The current iteration of the program, the Legged
Platform (TBCP-II), can move from horizontal to vertical
Squad Support System (LS3) (above) seeks to demon-
surfaces and over both inside and outside corners, as can
strate that a highly mobile, semiautonomous legged
other gecko-inspired climbing robots.7
robot can carry 400 lbs. of load through rugged terrain. 6

C. Scientists also are studying the way D. Harvard researchers are pushing the technology on
birds and insects fly to better an even smaller scale with their RoboBees project
understand aerodynamics and (above), which emulates a colony of honeybees
low-noise systems. The Nano containing insect-sized flapping-wing robotic platforms.
Hummingbird (above) demonstrated These biomimetic platforms have the added advantage
controlled, precision hovering and of stealth.9
fast-forward flight of a two-wing,
flapping-wing aircraft that carries its
own energy source and uses only its
wings for propulsion and control.8

24  Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)

 
Intelligence of the Future?

uncertainties of the IC’s real-world In all conditions they must exhibit formats that satisfy the user’s needs.
environments and tasks. endurance while maintaining sta- At some level, however abstract, a
ble operations, despite attacks and human-machine interface will be
Current technologies may have the challenges to propulsion, sensors, necessary, even when autonomous,
potential to permit greater autonomy, and communications. Speed, agili- humanlike decisionmaking may be
but they primarily now work with ty, and stealth are essential parts of more sophisticated than it is today.
well-defined rules and exhibit only the IC mission and are ambitious
limited autonomy in initiating and requirements when combined with Analysis and Decisionmaking
carrying out innovative tasks intelli- autonomy. Today’s user-system interfac-
gently. Surveys done by the Interna- es place the burden on humans to
tional Federation of Robotics (IFR) Ideally, an autonomous system acquire relevant data. As described
Statistical Department show that should be able to make informed in Psychology of Intelligence Anal-
there have been worldwide increases decisions with human guidance. ysis—a frequently cited document
of industrial robots and greater use Current systems require programmers despite its age—IC analysts use
of service robots. Industrial functions to understand mission parameters and formal and less formal methods
include tending of machines (metal translate them into well-formed rules to identify data, draw inferences,
work and plastic molding), pallet- based on the end user’s understand- and find answers.11 In more formal
izing and inventory, and dispensing ing of the problem. analysis, inferences are made through
(painting, sealing, gluing). Service procedures that collectively repre-
robots work on tasks ranging from This approach allows the auton-
omous system to make decisions sent the scientific method, including
household chores (cleaning, lawn statistical analysis of data on the
mowing) to dirty, dull, distant, dan- but only in the context of how the
programmer described the problem. phenomenon in question. To make
gerous, or repetitive tasks.10 analytic arguments, assumptions, or
Truly autonomous systems must be
Autonomous system technologies able to do complex tasks and perform intelligence gaps more transparent,
must be proficient at operating in a missions in situations that require a variety of diagnostic techniques
variety of complex, dynamic environ- modifications to rules. This capability supplement these methods.
ments, all of which present challeng- either will come from expanding rule The goal of autonomous systems in
es. For example: sets or giving systems the ability to the realm of intelligence analysis is
learn from their environments. Such to provide, in combination with other
• In space: zero-gravity, airless, advances will require much more
extremes of cold and heat; data sources, an extensive, complex
research. database to assist with analytic tech-
• Air: variable pressures, winds, Humans Are Part of the System niques. How will analysts quickly
extreme and sudden weather extract information relevant to their
shifts; A human-autonomous technology decisions and narrow the possibilities
interface may seem like an oxymo- to significant or actionable items?
• Land: variable gradients or slopes, ron. Today, however, it is an essential
multitudes of obstacles, man- ingredient of semiautonomous sys- A critical foundation for this next
made and natural; tems under direct or indirect human generation of decision support is to
control. In the future, autonomous develop methods that automatically
• Sea: high seas, sharply variable systems must respond even more to respond to a decisionmaker’s needs.
currents, changing weather con- human needs and give meaningful One vision is a system that would
ditions; feedback. For example, a user may provide decisionmakers alternative
guide a sensor that is capable of courses of action by selecting, col-
• Undersea: changing pressures of
doing surveillance of all data within lecting, and formatting data relevant
depth, variable terrain near sea
its frequency and range limitations to to a problem. The purpose would
bottom, changing currents, and
focus on certain target characteristics. be to give decisionmakers infor-
changing temperatures.
The sensor system’s download capa- mation-rich environments in which
bilities should send relevant data in choices are clearly laid out and from

Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)  25


Intelligence of the Future?

which they can rapidly take action process information, and suggest integration for autonomous systems.
specifically suited for the problem. decisions. Most examples shown so Maturity and integration are differ-
(“agile” and “adaptive” in current far only tackle one component, such ent aspects of system complexity
business jargon). In this realm, much as locomotion, at a time and only in that have to be addressed. Maturity is
more research must be done before laboratory conditions. based on the length of time the tech-
the IC can gain the benefits of auton- nology has been developed and used
omous analytical methods. The graphic below illustrates this and relates to solving issues dealing
point. Along the bottom is a measure with the use of the technology. For
The Potential of Biotechnology of task/environmental complexity. On example, GPS location services is a
Future decisionmaking interfaces one end are very simple tasks such as mature technology, but automatic
will probably integrate advances sensing and moving along the ground identification of objects is much less
in biotechnology into autonomous in an open area. Laboratory envi- mature. Integration has to do with
systems. In a concept known as “aug- ronments are also at this end. Tasks the connection, communication, and
mented cognition,” real-time mea- such as flying, swimming, and pattern interaction of components to work
surements of chemical changes that recognition are more complex. IC together and function properly. Inte-
take place during human cognitive mission-level tasks such as “moni- gration differentiates between simple
processes may be used to provide tor this 10-mile stretch of border for systems, such as geolocation, and a
information to platforms so they can vehicles or people” or “patrol this complex system, such as goal-direct-
sense a human’s thinking and act area underwater and record all com- ed surveillance linked to multiple
accordingly. For example, orexin is a munications” are complex tasks that response options.
neurotransmitter that regulates arous- involve multiple system components,
al, wakefulness, and appetite.12 In the integrated and working together. These definitions help to differ-
past, orexin could be measured only entiate autonomous systems. While
The left side of the graphic is a a simple system can be capable of
with a spinal tap. Now, we can get measure of technology maturity and
the same data using measurements complex actions, a level of increased
in saliva.13 Some interfaces might
leverage noninvasive brain-machine
interface methods such as electro- HIGH
encephalography, which records
the small electric field generated by
Relative
groups of neurons firing, and near-in- Degree of Autonomy
frared spectroscopy, which uses Needed for IC Tasks
Maturity/Integration Required

scattered light to detect changes in

the brain.14

Analysis of Current Au-

tonomous Systems

Many examples exist of autono- Relative

mous systems being developed by the Degree of Autonomy
in Current Systems
military, industry, academia, com-
panies, and hobbyists. None has the
level of autonomy needed to handle LOW COMPLEX
the missions that most interest the Task/Environmental Complexity
SIMPLE
IC. To complete complex tasks, a
system needs to do many things, such
as interact with its environments,

26  Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)

 
Intelligence of the Future?

maturity and integration can improve omous systems,15 but it is considered past that to using service robotics as
the robustness and range of system to be in the early stages of competing a solution for looming societal and
capabilities. Integration also informs with robot pioneers in other coun- demographic problems.
the number of independent com- tries, including Japan, Switzerland,
ponents that a system has. For the Germany, Sweden, and the United Humanoid robots are a particular
examples mentioned in the previ- States.16 China’s 15-year science and area of new development and include
ous sections, the levels of maturity technology plan published in 2008 auditory analysis and speech com-
are low. In some cases, the level identifies 22 specific technologies on munication, sensory integration, and
of integration is not there because which work is planned. Among these brainlike information processing.19
the technology is a component that are “intelligent perception technolo- However, its near-term relevance to
would have to be integrated into a gy” in the information sciences and the IC is difficult to derive other than
larger system to complete an IC task. “intelligent service robots” in the as interesting technology devel-
The graphic points to the amount of area of advanced manufacturing. opments in sensing, deciding, and
development needed to take today’s acting.
systems and demonstration systems The focus on industrial and
service robots is supplemented by Israel. An American Technion So-
to the appropriate levels of maturity
other developments such as a robot ciety (ATS) web article describes the
and integration required to complete
dolphin that swims through water to scope of Israel’s interests in autono-
IC tasks successfully.
measure its quality and robots that mous systems as shown in a program
exhibit capabilities for more complex of the Technion-Israel Institute of
tasks such as assisting with surgical Technology.
The Known State of De- procedures. Two leading robotics
The program is described as a
velopment Abroad laboratories in China have computer
“multi-pronged” approach to surveil-
science research including pure artifi-
lance, including the use of swarms
Several countries are known to be cial intelligence (AI) that specializes
of UAVs and/or satellites with the
investing in technologies to advance in natural language, machine trans-
capability to cooperate with one an-
autonomous systems, but publicly lation, and reinforcement learning,
other while operating under extreme
available information suggests the with related AI work in multimedia,
conditions on the ground and in the
United States is the most advanced distributed computing, and pervasive
sea. The program even includes ex-
in the field. Its systems are capable computing.
perimentation, the article claims, on
of working in more complex envi-
European Union. A European “bio-inspired” snake robots that pro-
ronments and at more complex tasks
Commission report, Future and pel themselves segment by segment
than systems being developed in
Emerging Technologies (FET) Proj- as a snake does.20
Japan, Korea, China, and Israel. All
three countries have made significant ects Compendium for 2007–201217
Turkey has begun to produce its
investments in industrial robots to presents research under an informa-
own surveillance and attack capabili-
improve manufacturing processes tion and communications technol-
ties as a move away from depending
and in service robots to help meet the ogy program, including autonomy
on other countries for such tech-
needs of aging populations. Assess- initiatives such as adaptive autonomy
nology. Its procurement authorities
ments of work in the European Union and autonomy of collectives such as
are preparing to field a semiauton-
and Turkey place their efforts behind robot swarms.
omous surveillance aircraft capable
China and Japan’s. of munitions delivery. The Turkish
Japan and Korea embrace the
power and potential of robots and government announced in March
The following are snapshots based
autonomy, particularly as applied to 2013 that it plans to sign a contract
on open sources of autonomous sys-
the industrial and service industries.18 for the acquisition of 10 locally made
tem research in other countries.
Industrial robotics already supports drone systems, dubbed the ANKA.
China seems poised for a vigorous export-driven manufacturing indus- One version will be armed. Turkish
and continuing investment in auton- tries and is, as in China, looking officials often look pleased portraying
the ANKA as a “fully Turkish, nation-

Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)  27


Intelligence of the Future?

al, purely indigenous aircraft,” but the standing, and multiagent coordina- ples. For example, flying nanorobots
drone’s imported parts include the en- tion. And even with developments like the RoboBees pictured earlier
gine, automatic take-off and landing in these areas, fully autonomous are inspired by biological research
system, landing gear, and radio.21 systems will only be achieved and build on the known behavior
through the integration of these and of insects. Inventions are often the
other technologies—some still to be products of young adults with fertile
invented (including hardware and minds or are creations designed to
Conclusions software). These developments will meet special human needs.22 They
have to come from at least three can be inspired by commercial tech-
Autonomous systems are tools sources, none of which is likely to nologies (e.g., toys, hobby kits, and
intended to increase the proficiency alone provide all the needed innova- computers) and vivid imaginations.
of key capabilities. In the case of tion, invention, and applications: Independent investment of resources
intelligence, they would help gather, is the engine of change when the in-
analyze, and protect vital informa- • academia (journals, university novation or invention shows promise.
tion, but current systems are not close projects);
to the levels of maturity and systems Today, except in the field of indus-
integration that would allow them to • inventions (toys, hobby kits, so- trial robotics, autonomous systems
autonomously undertake complex IC cial intelligence devices); and exist at the level of innovation and
missions. invention. Even the semiautonomous
• independent technology invest-
robots already developed for military
Advances in technologies will be ments.
and security purposes (e.g., iRobot;
needed to achieve such capabilities, L3 CyTerra) for surveillance and
Academia always has been a
especially in the fields of perception, aerial and ground delivery of supplies
source of innovation and new devel-
planning, learning, human-robot and equipment are limited production
opment. It tests innovative directions
interaction, natural language under- items.
by building on theories and princi-

v v v

28  Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)

 
Intelligence of the Future?

Endnotes:

1. Defense Science Board, Task Force Report: The Role of Autonomy in DoD Systems (Office of the Under Secretary of Defense for Acqui-
sition, Technology and Logistics, July 2012), 15 (available at http://www.fas.org/irp/agency/dod/dsb/autonomy.pdf.)
2. https://www.cia.gov/about-cia/cia-museum/experience-the-collection/index.html#!/ artifact/14 
3. https://www.cia.gov/about-cia/cia-museum/experience-the-collection/index.html#!/ artifact/16
4. Hugh Durrant-Whyte and Tim Bailey, “Simultaneous Localization and Mapping: Part I,” Robotics and Automation Magazine 13, no.
2 (2006): 99–110, at http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=1638022; Tim Bailey and Hugh Durrant-Whyte, “Si-
multaneous Localization and Mapping (SLAM): Part II,” Robotics and Automation Magazine 13, no. 3 (2006):108–117, doi: 10.1109/
MRA.2006.1678144.
5. See “Scalable sWarms of Autonomous Robots and Mobile Sensors (SWARMS) Project,” http://www.swarms.org/.
6. See “BigDog –The Most Advanced Rough-Terrain Robot on Earth,” http://www.bostondynamics.com/robot_bigdog.html; See “Legged
Squad Support System (LS3),” http://www.darpa.mil/Our_Work/TTO/Programs/Legged_Squad_Support_System_(LS3).aspx.
7. J. Krahn, Y. Liu, A. Sadeghi, and C. Menon, “A Tailless Timing Belt Climbing Platform Utilizing Dry Adhesives with Mushroom Caps,”
Smart Materials and Structures 20, no. 1 (2011): 115021–115031, doi: 10.1088/0964-1726/20/11/115021; TaeWon Seo and Metin Sitti,
“Tank-Like Module-Based Climbing Robot Using Passive Compliant Joints,” IEEE/ASME Transactions on Mechatronics 18, no. 1
(February 2013): 397–408, doi: 10.1109/TMECH.2011.2182617; See “Waalbot: A Wheel/Leg Robot for Wall Climbing,” http://nano-
lab.me.cmu.edu/projects/waalbots/tri-leg.shtml.
8. See “Nano Hummingbird,” http://www.avinc.com/nano.
9. See “RoboBees: A Convergence of Body, Brain, and Colony,” http://robobees.seas.harvard.edu/.
10. See “Welcome to the IFR Press Conference, 30 August 2012, Taipei,” http://www.worldrobotics.org/uploads/tx_zeifr/Charts_IFR__30_
August_2012.pdf.
11. Richards J. Heuer, Jr., Psychology of Intelligence Analysis (Central Intelligence Agency: Center for the Study of Intelligence, 1999),
https://www.cia.gov/library/center-for-the-study-of-intelligence/csi-publications/books-and-monographs/psychology-of-intelli-
gence-analysis/PsychofIntelNew.pdf.
12. Sere “Orexin,” http://en.wikipedia.org/wiki/Orexin.
13. J. Hagen, W. Lyon, Y. Chushak, M. Tomczak, R. Naik, M. Stone, and N. Kelley-Loughnane, “Detection of Orexin A Neuropeptide in
Biological Fluids Using a Zinc Oxide Field Effect Transistor,” ACS Chemical Neuroscience 4, no. 3 (March 20, 2013): 444–453, doi:
10.1021/cn300159e.
14. E. Akhadov, A. Speed, A. Konkel, and R. Held, The Human Side of the Net (Sandia National Laboratories, Sandia Systems Assessment
and Research Center, March 2010).
15. David Atkinson, Assessing Current Research on Autonomous Systems and Related Technologies in Australasia (Institute for Human
and Machine Cognition, August 1, 2011). http://www.academia.edu/1056646/Assessing_Current_Research_on_Autonomous_Systems_
and_Related_Technologies_in_Australasia.
16. See “China Looking to Lead on Robot Innovation,” http://www.southerninnovator.org/index.php/innovation/196.
17. Future and Emerging Technologies (FET) Projects Compendium 2007-2012. Commission of the European Union, Communications
Networks, Content & Technology Directorate-General, Future and Emerging Technologies unit.
18. IACF Industry Study Group, Robotics and Autonomous Systems Industry (National Defense University: Industrial College of the
Armed Forces, 2011), http://www.ndu.edu/es/programs/academic/industry/reports/2011/pdf/icaf-is-report-robotics-autonomous-sys-
tems-2011.pdf.
19. Ibid, 15.
20. See “The Technion Autonomous System Program (TASP),” http://www.ats.org/site/PageServer?pagename=about_research_autono-
mous_systems.
21. Turkish UAV market may boom after development Burak BEKDİLHürriyet Daily News, August 27 2013
22. Harry S. Dent, The Great Crash Ahead: Strategies for a World Turned Upside Down (Free Press, 2011).

v v v

Studies in Intelligence Vol 59, No. 1 (Extracts, March2015)  29