The interplay between the human and objects in
ISR practices and targeted killing operations
A critical look into how the role of the human is reconfigured in current automated Intelligence, Surveillance and Reconnaissance (ISR) practices and targeted killing
operations conducted by Unmanned Aerial Vehicles (UAV’s)
Master thesis International Relations (Political Science) University of Amsterdam, 20-06-2016
Student: Louk Bracco Gartner (10829628) Supervisor: Dr. P.E. (Polly) Pallister-Wilkins Second reader: Prof. dr. M. (Marieke) de Goede
2
Abstract
Intelligence agencies and governments currently have to work in a security environment in which the automation of Intelligence, Surveillance and Reconnaissance (ISR) practices prevails, and execution capabilities of Unmanned Aerial Vehicles (UAV’s) are at the light of day. Both media reports and scientific scholarship are increasingly focusing on the agency of technology in ISR practices and execution capacities of UAV’s, but the consequences of these developments on the role of the human in these practices remains a topic that is not examined thoroughly.
This thesis uses a qualitative (inductive) and critical approach in examining the role of the human compared to technology, and rather raises questions on current developments than it aims to answer them. Bruno Latour’s concept of “dingpolitik” and Eyal Weizman’s concept of “forensic architecture” will provide the theoretical background that illustrates the material agency of automated ISR practices and execution capacities of UAV’s. It will present a framework of how objects, in all its facets, have an increasing impact on our socio-political world.
After interviews with experts in the fields of intelligence, surveillance, terrorism and technology, four themes were developed to examine the reconfigured role of the human in automated ISR practices and targeted killing operations: the uniqueness of human action in both observation processes and interpretation practices, the lethal impact of machines on the lives of humans, and the issue of accountability in ISR practices and targeted killing operations. These findings will present a critical look into, and add new insights to, the interplay between humans and objects in security studies.
3
Contents
1. Introduction 4
2. Human and technology in social sciences 6
2.1 Materiality matters 7
2.2 RMA and intelligence gathering methods 11
2.3 The role of the human in current ISR practices 13
3. Methodology 15
4. The reconfiguration of the human in ISR practices and 16 targeted killing operations
4.1 The uniqueness of a human touch in observation processes 19 4.2 The uniqueness of a human touch in analysis practices 23 4.3 The lethal impact of machines on the lives of humans 29 4.4 Accountability in ISR practices and targeted killing operations 31
5. Conclusion 36
4
Introduction
The United States have allegedly engaged in drone strikes outside of a declared war zone since 2002, but it is only in 2013 that the White House releases a set of standards, procedures and guidelines for conducting their strikes (Scahill, 2016: 2-3), and the outside world was able to look into these infamous practices. These strikes are conducted by Unmanned Aerial Vehicles (UAV’s), or drones, and in different media and academic articles are also referred to as “targeted killings”, “signature strikes”, or “extrajudicial killings” (Klaidman, 2012; Scahill, 2016; Shaw; 2013). The definitions vary in essence not much, and will not be thoroughly discussed in this thesis, but they all encompass the killing of suspected terrorists in an undeclared war zone based on mostly technical intelligence (ibid).
This development is part of a more abstract discussion that will be presented in this thesis. The focus will lie on how innovations in technology are responsible for the automation of Intelligence, Surveillance, and Reconnaissance (ISR) practices and advent of UAV’s (capable of conducting targeted killings), and how this development has an effect on the role of the human in different phases of these processes.
The starting point of this thesis is therefore that intelligence agencies and governments currently have to work in a security environment in which the automation of ISR practices prevails, and execution capabilities of UAV’s are at the light of day. Scahill (2016: 5, 110-120), an American journalist and author of several books on drone strikes, cites in his most recent book the observations of former employees of U.S. intelligence agencies and military departments on the overreliance of the military on signals intelligence (SIGINT): “Taken together, the secret documents lead to the conclusion that Washington’s fourteen-year-long high-value targeting campaign suffers from an overreliance on signals intelligence, an assassination rather than capture, an inability to extract potentially valuable intelligence from terror suspects.” This turn from a human-oriented approach towards a more material approach in this particular field of security studies will be the main focus of this thesis.
Both media reports and scientific scholarship are increasingly focusing on the agency of technology in ISR practices and execution capacities of UAV’s, but the consequences of these developments on the role of the human in these practices remains a topic that is not examined thoroughly. Critical writers focusing on the use of UAV’s in surveillance and execution practices are worried about the automation of security practices, the subsequent declining role of the human, and the ethical questions this raises (Heller, 2013; Kindervater; 2016; Klaidman; 2012; Shaw; 2013).
5
This research will use a critical research design that aims to open up inquiry. Its goal is to put forward relevant questions, rather than answering them, in order to explore and critically assess the socially constructed world we now live in (Mutlu & Salter, 2012: 31). Therefore, it will approach the increasing use of UAV’s in surveillance and targeted killing practices from a material point of view. Mutlu (2012: 173) elaborates in his work on the scholarship that studies objects beyond their material existence and looks at their social and political impact on, for example, humans. The automated ISR practices, and executing mechanisms of UAV’s in targeted killing practices, are in this thesis the objects that have material agency.
However, the changing role of the human in these practices remains indeterminate. Critical questions are to be raised, since UAV’s and subsequent software now have the capability to observe, analyse and possibly carry decision-making authority, whereas these practices were usually reserved for humans (Kindervater; 2016). These processes of observation and monitoring, the analysis by algorithms, and decision-making capacities (based on software, and on a political level) are the three main levels of inquiry in this thesis. How did the role of the human change in each of these stages? What are the consequences of these changes? Has the value of human input in these practices declined, increased or did the focus only shift? What is the role of politics in these ISR practices? These are some of the questions this thesis tries to answer, using insights from intelligence and security experts and former employees of intelligence agencies, in order to provide an answer to the central research question: How does the role of the human get reconfigured in current automated Intelligence, Surveillance and Reconnaissance (ISR) practices and subsequent targeted killing operations conducted by UAV’s?
The data presented in this thesis were obtained by expert interviewing, in order to elaborate on a more philosophical level on the role of the human in an area of security studies that is largely focused on the use of technology. By presenting their insights on the changing role of the human in the aforementioned three stages, this thesis aims to critically explore the value of human actions in current automated ISR practices.
This thesis will be organized as follows. The theoretical section will focus on how materiality matters. The material agency of UAV’s will be set out in order to present a framework of how technology, signals intelligence (SIGINT) and software have an increasing impact on our socio-political world. Human testimony is regarded obsolete, whereas the interpretation of materials is exclusively assigned to technology or so-called experts. This will be illustrated with the use Bruno Latour’s concept of “dingpolitik” (2014) and Eyal
6
Weizman’s (2011) concept of “forensic architecture”. In the methodology section the process of expert interviewing and its implications will be presented. The analysis will elaborate on how the role of the human is reconfigured in the aforementioned three stages, using four themes that recurred during the interviews: the uniqueness of human action in both observation processes and interpretation practices, the lethal impact of machines on the lives of humans, and the issue of accountability in ISR practices and targeted killing operations. The conclusion will answer the central research question, present the implications of the results on existing theory and literature, address the shortcomings of this research, and make suggestions for future research.
2. Human and technology in social sciences
The configuration of the ‘human’ in comparison to, and influenced by, technology is a subject that is highly debated in different fields of social sciences. Humankind has always been exposed to different aspects of life and technology is just one of the main developments that had an increasing influence on the reconfiguration of the human being. In academic research many scientists try to identify these changes and the role of technology in this matter. A clear example of this is the research done by Law (1984: 234) on methods of long-distance control by the Portuguese in the late 15th century and early 16th century. The Portuguese wanted to equip their vessels with navigational context that was less dependent on European geography and more focused on durable and forceful equipment in order to obtain global mobility. Law contends the combination of documents, devices and drilled people were the key to success, only when placed in the right structure. The device, or technology, Law refers to is the astrolabe. The Portuguese developed it into a form that was small and durable, facilitating them in their conquests around the globe (1984: 251-253). This deployment of technology was partly due to the unreliability and non-durable state of the human (1984: 254-258).
Another interesting perspective on alternatives of human action is provided by Cudworth & Hobden, who contend there is a ‘posthuman way of war’ that urges the field of international relations and security studies to critically assess the deeply human-centred character in almost all academic articles. Almost none of these studies focus on how actors other than humans can have an influence on the conduct of war. In order to capture more accurate systems of war it is important to acknowledge the role animals can play in the landscape of war. This line of research is related to feminist scholarship in a sense that it critically assesses state-centrism and reconfigures the role of the human (1-3, 12-13).
7
However, in this thesis the focus will lie on the configuration of the human compared to technology in security studies, and ISR practices in general. Before turning to the more specific concepts of human intelligence (HUMINT) and signals intelligence (SIGINT), the relationship between technology, the human, and politics will be discussed by elaborating on Bruno Latour’s (2014) concept of “dingpolitik” and Eyal Weizman’s (2011) “forensic architecture”. These concepts will provide clear insights in the expanding role non-human aspects (e.g. material, technology) play in ISR practices, significantly influencing the role of the human in these practices.
2.1 Materiality matters
According to Walters (2014), the theoretical framework of dingpolitik – a concept introduced by Latour (2005) - can provide security studies with insights of the comprehensive relationship between technics and politics. Materiality matters, and therefore ordinary material objects – not just technological objects – can influence and constrain security governance (2014: 102). Material things can become entangled in disputes and therefore offer important insights about the relationship between technics, politics and the public.
For Latour (2005: 4) realpolitik is “a positive, materialist, no-nonsense, interest only, matter-of-fact way of dealing with naked power relations”, whereas his concept of dingpolitik illustrates how objects and things can be fleeting, ambiguous, partial and more-than-single (Latour, 2005; Walters, 2014: 103). Dingpolitik is concerned with politics oriented around disputatious “matters of concern” (Walters, 2014: 104, Latour, 2005: 13). The whole idea of indisputable facts is impossible and political controversies always exist in a scheme of representation (Walters, 2014: 104; Merleau-Ponty, 1968).
Latour (2005: 4) uses an “object-oriented democracy” in which dingpolitik identifies multiple assemblies. He contends that each object – each issue we care about or reflect as a matter of concern to us – generates different patterns of emotion, disruption, and agreement or disagreement (2005: 5). No coherence or continuity might be found in our opinions, but there is hidden continuity in the things we are attached to: “Each object gathers around itself a different assembly of relevant parties” (ibid). Before Latour published his work, other authors such as Harré (2002) already paid attention to the phenomenon of objects and their meaning. He exemplifies his insights on this topic with fairy tales and how objects in these folk tales carry narratives with them. He defines material things as having “magic powers only in the contexts of the narratives in which they are embedded” (2002: 25). With ‘magic power’ he
8
means a power that is not an effect of the physical properties of the object in question, rather, he contends that the power of a form of words is what configures this magic aspect (ibid).
Leander (2013) also elaborates on Latour’s work, but uses the concept of technology instead of objects. Which, in principle, is something different, but both are non-human alternatives that generate meaning and have a profound impact on political disputes. She cites in her theoretical framework how “technologies and things more generally ‘can authorize, make possible, encourage, make available, allow, suggest influence, hinder, prohibit and so on’. They have performative effects that escape human intention and control” (2013: 814-815). In other words: technology has material agency. Another recurring idea she mentions is that technologies generate meaning. Technologies are capable of generating classifications, and ways of interpreting. She contends, in line with Latour’s observations, that technologies are at the origin of interpretations and representations of things. She explicitly states that things/machines should not be considered human. On the contrary, “material should be understood on its own terms and in context” (2013: 816).
Thus, dingpolitik sensitizes how subjects and things can become active participants in political disputes (Walters, 2014: 104), and Walters exemplifies this with a study on a report from the international human rights organization Human Rights Watch (HRW) entitled Precisely Wrong: Gaza Civilians Killed by Israeli Drone-launched Missiles (2009). He shows how the mobilization of objects and materials is essential for how NGO’s try to bring visibility to any injustice in this unstable region of the world. This report exemplifies very clear how making things (objects) public, as Latour and others suggested in their articles, can have a significant influence on political disputes (Walters, 2014: 107-108).
Before going more in-depth on how this material agency works, it is important to understand that objects never ‘speak’ without assistance (2014: 109). The human is still responsible for observing and interpreting the object and formulating a certain representation. The Precisely Wrong report provides a clear understanding of whose responsibility it is to speak in their name. These spokespersons are almost exclusively military specialists or forensics experts (ibid). According to HRW they are experts in their field of work and capable of objectively analysing objects.
Walters (2014: 104) comments on the report by stating that the “HRW accorded forensic knowledge about material objects a more prominent and more privileged place in its case against the IDF than the voice of the victims themselves”. To elucidate this line of reasoning he believes Eyal Weizman’s (2011) concept of “forensic architecture” - a useful
9
mid-range level of analysis that is closely related to the more abstract concept of dingpolitik – can provide clearer insights in the role of materials in political disputes.
Weizman (2011) introduces this concept in his book ‘The Least of All Possible Evils: Humanitarian violence from Arendt to Gaza’. He uses the same HRW report as mentioned above and describes how the HRW report focused largely on the examination of material remains and on the military technology being used. His definition of the “forensic architecture” is as follows:
“Much of this research was concerned with the ‘interrogation’ of the ruins and rubble of destroyed buildings, (…). The emphasis on the investigation of material remains and their overwhelming abundance meant that the forensic analysis of built structures – a practice I would like to refer to as ‘forensic architecture’ – emerged at the forefront of the legal– political disputes that ensued. Existing at the intersection of architecture, history and the laws of war, forensic architecture must refer to an analytical method for reconstructing scenes of violence as they are inscribed in spatial artefacts and in built environments.” (Weizman, 2011: 100-101).
Weizman (2011) also adduces Richard Goldstone’s report on the same topic, named Report of the United Nations Fact Finding Mission on the Gaza Conflict (2009). The interesting part of this research comes to the surface in the methodology section. Goldstone was considered a proponent of oral testimonies in earlier investigations he led, but in the case of the Gaza Conflict report a shift of emphasis from human testimony to material evidence took place (Weizman, 2011: 103). Even though the methodology section implies that there are dozens of human testimonies involved, the balance tips over towards objects and their objectivity. The reliability of human testimony is contested, and therefore the focus in the research methods was on analysis of video and photographic images, satellite photographs, expert analysis on such images, medical reports, weapons and ammunition remnants, and meetings with analysts, medical doctors, scientists etc. (Weizman, 2011: 103, Goldstone, 2009: 14-16). This switch of emphasis is a development Weizman (2011: 103-104) refers to as “forensic fetishism”; a subcategory of the broader concept of “forensic architecture”. He contends that forensic analysis of this kind provides scientific evidence pronounced by expert witnesses. Since humans are capable of lying, much more value is put on non-human evidence that speaks for itself. The testimony of victims is considered unreliable, and thus the difference between “people and objects is that subjects can misremember or skew their testimony in
10
relation to their political self-interest while an evidentiary truth seems to linger, fossilized in the object, ready to be unpacked by science” (ibid).
Weizman then turns to classical Rome to illustrate how forensics always have consisted of both human speech and objects and are part of a total forensic rhetoric (2011: 105). Since objects do not speak for themselves, there is always need for translation or interpretation. This requires a human being. Nowadays this role is appointed to the so-called experts. Therefore, it is very important that objects are presented in a non-biased way, in their most authentic condition (ibid). But that is the hard part, since there is always rhetoric involved. Weizman (2011: 107) describes this rhetoric as follows: “the gestures, techniques and technologies of demonstration, methods of theatricality, narrative and dramatization”.
This averting of biased acting by humans is something Johnson (2013) addresses in combination with politicization. He contends that the goal of intelligence agencies is to find conscientious intelligence professionals dedicated to empirical fact-finding and unbiased assessment of facts (2003: 655). The politicization of information – meaning the bending of intelligence reports to a consisting political position from policy makers - is one of the biggest threats to independent reporting on any kind of subject (2003: 654-655). Therefore, he looks very critically at the most significant inadequacies of intelligence, such as the lack of timeliness, the frequent unwillingness of policymakers to accept reliable information and the risks posed to democracy by the politicization of information (ibid).
Holmqvist (2013: 544-545) too takes into account the agency of material objects and their influence on politics. He argues that drones are political actors in the sense that they are embedded within the imperial and military apparatus behind them: “The ascription of human relations to the steely construction of the drone is borne out in analysis of how the drone could not have been invented or constructed other than as part of the particular apparatus of military power that envisaged their use” (2013: 545).
These insights, especially those provided by Weizman, show how material has agency, but it also exemplifies how humans will always have an important role to play in observation and interpretation processes. Weizman’s conclusion is that this “era of forensics” (2011: 115) is overcompensating the indeterminate and fragile voice of the victim (the human). The tilt towards the interpreted “material witness” is not a fixed alternative for the subjective human testimony (2011: 114). Rather, the problems associated with human testimony come back in current forensic practices as well.
11
2.2 RMA and intelligence gathering methods
This interplay between the human and technology (and materials) is especially rendered visible in ISR practices. Before turning to more recent discussions on the role of the human compared to technology and materials it is important to conceptualize both the advent of recent technological innovations and the two main intelligence gathering methods used in this paper: HUMINT and SIGINT.
Since the turn of the century the amount and quality of ISR practices increased significantly. Rasmussen (2006), among other scientists, considers the ‘Revolution in Military Affairs’ (RMA) as the main cause that set this development in motion. He argues that an acceleration in technological innovations took place in the 1990’s and ever since unprecedentedly reshaped and influenced the future of warfare. The RMA provided a new rationale for how the development of information technology had an influence on Western strategic discourse. In the final years of the Clinton administration and the early years of George W. Bush’s the nature of warfare was to be defined in a universe of continuous transformation where means and ends are intertwined (2006: 49). The classical means-end rational Clausewitzian thinking of how means are related to ends was considered obsolete. Mastering technological means had become an end in itself (2006: 45). Mainly, because the RMA would be able to provide near-perfect information on different aspects of war and therefore remove the ‘irrationalities’ of terrain, weather and human horror (2006: 53). Thus, generals and politicians started to use the RMA as a promotion tool for their political goals, because of the seemingly endless new possibilities these innovations in information and communication technologies created. Therefore, technology became the pivotal aspect of modern warfare (Rasmussen, 2006: 43-52).
Before going into more detail about the current status of ISR practices, and the role of technology and the human in this undertaking, it is important to conceptualize both SIGINT and HUMINT practices. Especially since this thesis will focus extensively on the observation and analysis (interpretation) part of intelligence gathering. The US Department of Defense Dictionary of Military and Associated Terms (2016) defines human intelligence (HUMINT) as “a category of intelligence derived from information collected and provided by human sources”. At first sight a rather narrow description for something of such importance for homeland security, but in fact it covers the most important aspects of intelligence gathering and surveillance practices done by a human being: observation and interpretation.
Margolis (2013: 45) amplifies the concept of HUMINT by stating that it “requires a great deal of time and resources to gather assets and analyse information, rendering it one of
12
the most difficult types of intelligence to produce and implement.” Nevertheless, according to Lowenthal (2014: 102-110) “HUMINT is far less expensive than the various technical collectors, although it still involves costs for training, special equipment, and the accoutrements clandestine officers need to build successful cover stories.”
In line with Weizman’s description of “forensic architecture” Margolis (2013: 46) acknowledges the value of agents as experts in their respective fields, and the possibility they provide of watching developments unfold first-hand. Arthur Hulnick (1999: 36) states that “agents can bring us material that cannot be obtained by technical sensors or developed by diplomats. It seems foolish to give up the possibility of learning inside information from a well-placed source.” It is important, however, to understand the susceptibility of HUMINT to deceptive tactics based upon counterintelligence (CI). For example, foreign states can be aware of the identity of an informant and force them to divulge wrong or misleading information to their superiors. Also, the usage of volunteers, or walk-ins, by intelligence agencies can be very risky, since you do not know as an agency if this person is reliable or trustworthy (Margolis, 2013: 46). Further use of the concept of HUMINT in this thesis will entail the observation and interpretation in the process of intelligence gathering done by a human.
According to Michael Turner (2014: 234) signals intelligence (SIGINT) is “the interception and decoding of foreign electronic communications, traffic analysis, and monitoring of friendly communications for security reasons.” It is made up of subsidiary collection disciplines, such as communications intelligence (COMINT), the interception of electronic emissions; electronic intelligence (ELINT) and telemetry intelligence (TELINT), which detects signals given off from weapons (Margolis, 2013: 46-47; Turner, 2014: 234).
The RMA and subsequent technological advances made it possible for several states - and especially for the United States - to increase their SIGINT operations worldwide. It is argued that the age old practice of espionage became an afterthought because of this development (Margolis, 2013: 43). Lowenthal (2014: 96) contends with regard to SIGINT: “It gives insight into what is being said, planned, and considered. It comes as close as one can, from a distance, to reading the other side’s mind.”
Another form of intelligence gathering is called geospatial intelligence (GEOINT), which Turner (2014: 113) defines as “animated renditions of imagery and geospatial data, which allow users to visualize inaccessible terrain”. According to Margolis (2013: 48) the United States started already in the 1950’s with U-2 flights to gather intelligence of Soviet defences. In the last decades Unmanned Aerial Vehicles (UAV’s) have taken over this role
13
since they are capable of providing hours of video (or satellite images) from, for example, dangerous terrain in Iraq and Afghanistan. The material evidence Weizman (2011) refers to in his explanation of the “forensic architecture” is, to a large extent, obtained by GEOINT practices.
Further use of the concept of SIGINT in this thesis will, for the ease of use, also entail GEOINT and other technical INT’s, since these forms of intelligence gathering are in their observation and analysing processes to a large extent dependent on technological means and materials. This juxtaposition of both of HUMINT and SIGINT should therefore function as a mid-range level of analysis on the role of the human compared to technology in security studies.
2.3 The role of the human in current ISR practices
There is a lot of recent literature on the, at first sight, declining role of the human and the forthcoming reliability on technological surveillance practices and, for example, the use of Unmanned Aerial Vehicles (UAV’s) in current warfare practices. Margolis (2013) argues in her article that some of the worst intelligence failures by the CIA of the 20th century can be attributed to a misguided trust on technology and an ignorance of human intelligence. Due to the simple fact that HUMINT can provide information current UAV’s are not capable of gathering. Also, she argues that an increase of unintentional civilian casualties are the result of the absence of HUMINT, and that may turn the tide of public support against UAV practices in time (2013: 56).
Read, Taithe & Mac Ginty (2016) also elaborate on the long pedigree between humans and their pursuit of technological innovations, and conclude that nowadays new technologies tend to focus more on technocratic questions of efficacy than on the human (2016: 5). A very important notion they take is that data is not knowledge, nor do the technological innovations have the capacity to analyse these data (2016: 2). In their research, the authors focused on the utility of data for humanitarian actors and their main finding was that they cannot cope with the amount of data they gather. On the one hand they are in possession of great information-gathering capabilities, but on the other hand lack the capacity to efficiently process the acquired data. The ideal of presenting a comprehensive, or ‘real’, picture of a security setting is seemingly impossible (2016: 10-13).
The fallibility of the human is a phenomenon that usually causes inconvenience, but in the case of weapons it can have a catastrophic effect. In the 1970’s Dumas (1980) already
14
wrote about the strive for perfection by humans in warfare and the impossibility of achieving it:
“The more weapons we deploy, and the greater their geographic dispersion, the more people will be interacting with them. And the greater will be the likelihood of disaster resulting from human error. Nor can we circumvent this dilemma by turning control over to machines, by somehow automating the human element out of the nuclear forces. For who designs the machines and who will build them?” (Dumas, 1980: 20).
He gives several examples about how the judgment of human beings can, on the one hand cause disasters, but on the other hand could also save the day: “We cannot escape our fallibility, so we must exercise the wisdom and the instinct for survival that are also fundamental of our humanity. We must find a way to coexist permanently with our innate imperfection” (ibid).
This reliance on objects, technology or other non-human instruments takes a very prominent position in current Intelligence, Surveillance and Reconnaissance practices (ISR). Especially with regard to targeted killing practices by Western states. Kindervater (2016) refers to today’s practices as “lethal surveillance”: “A practice in which mechanisms of surveillance and knowledge production and decisions on life and death have become one and the same” (2016: 2). The ability of today’s drone is not only to minimize risk by removing the physical presence of the human from war, rather, its significance lies in how the history of ISR practices (information gathering and processing) and lethal targeting have become intertwined (ibid). The Predator drone is no longer considered a ‘nonhuman’ alternative, but instead it provides a ‘more than human’ capability (2016: 10). The ability to gather and process data with greater speed and analytical capabilities contributes considerably to the automation of war. This speeding up of war, the related information overload, and subsequent complexity of analysis mean that the human can no longer participate in war as he did before. And because of this, proponents of this automation are convinced that in the future machines will inevitably be better at identifying targets and will be able to make the actual decision to kill (ibid).
The abovementioned theory on human action and the increasing role of technology in ISR practices shows the changing role of the human. Scientists have different views on the current value and importance of the human, and its role in the future. It is interesting to philosophize about how the role of the human has changed due to technological innovations
15
and what this means for its future importance. Are drones reliable enough to observe and analyse possible threats? Will technology really take over decisions of life and death? Will the human really disappear from the process of intelligence gathering?
Therefore, in the upcoming data and analysis section insights will be presented on the reconfigured role of the human compared to technology on three closely related levels: observation in current ISR practices, subsequent analysis of the obtained data, and decision-making capabilities. This in order to illustrate how the material agency of technology, and objects (drones), has influence on how the role of the human is reconfigured in ISR practices and targeted killing operations. Critical questions are to be raised concerning the automation of surveillance, targeting and decision-making processes and several experts in the fields of intelligence, surveillance and terrorism studies shine their light on this trend and the pivotal role governments and other political bodies play in it.
3. Methodology
In order to critically explore the role of the human compared to technology in current ISR practices and targeted killing programs seven in-depth, semi-structured interviews with experts in the fields of intelligence, surveillance, terrorism and technology have been conducted. Since this thesis uses a qualitative (inductive) and critical approach in examining the role of the human compared to technology, and rather raises questions on current developments than it aims to answer them, the data presented are the unique views of people with different backgrounds in security studies.
Five of the seven interviews were conducted with academics in the abovementioned academic fields and were approached by email. The other two interviews were conducted with people who have experience in working for intelligence agencies (former employees) and came about due to snowball-sampling. Some of the academic interviewees provided contact information of these people or contacted them by themselves, before providing me access to any personal information. The balance tips over towards academics in this thesis, but, as Nader (1972) and Gusterson (1997) state in their articles on the merits and problems associated with “studying up”, it was difficult to find any experts or policy-makers in the field of security studies that were willing to talk on this subject. During the process of requesting interviews some of the approached people expressed their hesitance towards talking about SIGINT operations, and some of them feared of being exposed after saying certain things that could be traced back to them (this was for one interviewee a reason to decline my request to
16
record the conversation). Therefore, the interviewees are mentioned only by their job title and not by their names.
The duration of the interviews varied from 45 minutes to two hours in which the interviewer and interviewees tried to talk on a more abstract level about the interplay between the human and technology. Nevertheless, abstract does not mean superficial and therefore the aim was to let the interviewees philosophize on the reconfiguring role of the human. Not only compared to the past, but the current and future role of the human was discussed thoroughly as well. Close to none of the questions were predetermined, only the three abovementioned levels of inquiry had to be discussed per se, in order to create a conversation in which the interviewees could talk freely on many different aspects. The interviews were experienced as open and informal and all conducted in Dutch. The analysis of these data was done by thematic analysis in which four themes were developed. These themes will be discussed in the next chapter.
The data obtained by these expert interviews are supported by findings, observations and views published in already existing literature on intelligence methods, surveillance practices, drone warfare, and the interplay between intelligence agencies and national governments. This in order to illustrate certain events, the increasing reliance on SIGINT instead of HUMINT, the advent of UAV’s in surveillance and targeted killing practices, and the analysis methods used in current SIGINT practices.
4. The reconfiguration of the human in ISR practices and targeted killing operations Out of the results of the seven interviews four main themes emerge that show how the role of the human has changed in ISR practices and targeted killing operations: the uniqueness of human action in both observation processes and interpretation practices; the lethal impact of machines on the lives of humans; and the issue of accountability in ISR practices and targeted killing operations. These four themes will all address the issue of reliability, since this a recurrent subject in all these discussions. The three aforementioned stages that are followed in ISR practices and targeted killing operations - observation, analysis and decision-making capabilities – de facto often mingle, and therefore it is difficult to analyse them as single entities. The four themes are established to develop and structure the main findings that, in the end, will provide an answer to the central research question.
In order to present these findings, a short introduction into already existing literature on ISR practices, ‘Big Data’ collection and targeted killing operations conducted by UAV’s is
17
necessary. The theoretical section already elaborated on the differences between HUMINT and SIGINT in observation processes, so the next step is to outline the main developments in current analysis practices. As mentioned before, observation and analysis are practices that have very close ties and are to a large extent intertwined. In every process of observation, some sort of interpretation takes place, since every person perceives certain actions in a different way (Hammersley & Atkinson, 2007: 3-5). Interpretation is almost exclusively a normative exercise, which makes it subjective. Observation done by machines, or technology, is regarded more objective, since a machine has no personal history or biased perception (Weizman, 2011: 15-16).
The automation of intelligence and surveillance practices has increased the amount of data governments and intelligence agencies are capable of collecting (Lyon, 2014). Since the turn of century these SIGINT practices have taken place, but it is in June 2013 that the whole world gets introduced to the phenomenon of ‘Big Data’ gathering. The revelations of Edward Snowden, former employee of the National Security Agency (NSA) in the United States, show that the NSA (and therefore the U.S. government) engages in incredibly large scale monitoring of populations, and also reveals how they do this (Lyon, 2014: 2).
Boyd & Crawford (2012: 663) define these Big Data as follows: “the capacity to search, aggregate and cross-reference large data sets.” This aggregating of data happens in a variety of areas, such as consumer marketing, health care, urban policing and anti-terrorism (ibid). The focus in this thesis will lie on how these Big Data are interpreted in ISR practices, and moreover, whose responsibility this is or should be.
In the United States and the United Kingdom respectively the NSA and the Government Communications Headquarters (GCHQ) are the agencies responsible for most SIGINT activities, whereas in the Netherlands the Joint Sigint Cyber Unit (JSCU) is incorporated in the ‘Algemene Inlichtingen- en Veiligheidsdienst’ (AIVD) and the ‘Militaire Inlichtingen- en Veiligheidsdienst’ (MIVD). The Dutch agencies are, compared to their equivalents in the U.S. and U.K., relatively small and have a limited scope, but this has no implications for this study on how the role of the human has been reconfigured in the process of data analysis.
Due to the automation in ISR practices and subsequent advent of Big Data dumps, technology also had to come up with software that would be able to analyse all these results. These algorithms allow users to systematically extract, disclose, and analyse data and turn this into what agencies hope will be actionable data (Lyon, 2014: 3). They are reportedly used in the so-called “signature strikes” (Byman, 2013: 2), or “targeted killings” (Scahill, 2016),
18
conducted by primarily American drones. The advent of UAV’s has made it possible for the United States to execute targets without having “boots on the grounds” (Byman, 2013: 2).
The use of drones by the American government is not a particular new phenomenon, but since the ‘war on terror’ these practices have intensified (Klaidman, 2012; Shaw, 2013). The increased capacities of UAV’s have made it easier for drones to circulate for an extended time above certain territories, collect more accurate data, and it created the possibility of taking out targets with devastating missiles (Scahill, 2016; Shaw, 2013).
Normally, the decision to invade a certain country, or send in special forces, is made after certain information is collected and discussed amongst intelligence officers, their superiors and political leaders. These targeted killing conducted by American drones have made it possible to actually let this process of information gathering, interpreting and decision-making mingle (Chamayou, 2015; Scahill, 2016; Shaw, 2013).
According to certain media, for example The Bureau of Investigative Journalism (2015) and The Intercept (2015), this execution of targets is controversial. TBIJ tries to keep records of the amount of drone strikes in certain countries, the targets that are killed, as well as the innocent civilians, whereas the The Intercept distinguishes itself primarily by being a platform for whistle-blowers. A big concern for them is the way these targets are selected. In the last years several media and academics have reported on the use of the “disposition matrix”, a type of software responsible for putting people on the ‘kill lists’ of the U.S. (Cobain, 2013; Shaw, 2013). Cobain (2013) defines this software as follows: “(…) a sophisticated grid, mounted upon a database that is said to have been more than two years in the development, containing biographies of individuals believed to pose a threat to US interests, and their known or suspected locations, as well as a range of options for their disposal.” Shaw (2013: 540) goes on by stating that the possible targets are not “identified by staff in the CIA’s headquarters in Langley, Virginia”. These are observations Scahill (2016) produces evidence for in his latest book “The Assassination Complex”. He (2016: 96-106) states repeatedly that targets are selected on controversial metadata analysis and cell phone tracking technologies, or as one of his sources states: “We are not going after people – we’re going after their phones, in the hopes that the person on the other end of that missile is the bad guy” (2016: 98). These are developments that raise questions about human action, reliability, accountability and humanity. The upcoming chapters will elaborate on these themes and critically assess them.
19
4.1 The uniqueness of a human touch in observation processes
The first theme that will be addressed is the uniqueness of human action in observation processes in ISR practices and targeted killing operations. The automation of ISR practices is a topic referred to in many different studies (Buxton & Gong, 1995; Ehrhard, 2010; Kindervater, 2016). This trend started after World War II, when states realized the importance of ISR practices in warfare (Keegan, 2003). The Cold War is indicative of the beginning of electronic information gathering with a growth of SIGINT often in favour over HUMINT (Kindervater, 2016: 5).
To show how the role of the human is reconfigured due to the automation of intelligence practices, it is important to study the whole concept of HUMINT more in-depth, determining the unique aspects of this intelligence gathering method. A former intelligence officer for the ‘Binnenlandse Veiligheidsdienst’ (BVD) and AIVD explains in detail that HUMINT is more than just the collection and transmission of information to intelligence agencies:
“When I worked for the service I had to recruit people. I needed to ‘build up’ agents from scratch in order to let them infiltrate in certain organizations to provide me with information. This whole process can take a long time. Sometimes years before you can infiltrate an organisation. But the thing is: it is essential for an intelligence agency. These agents can provide you with information an automated drone will never find. Until this day, the recruitment of agents, and infiltration of organizations is the most important feature of an intelligence agency.”
These agents are not only recruited for domestic work. Nowadays these agents are more frequently used to infiltrate in organizations in foreign countries as well. Very important in this matter is therefore the cooperation with foreign intelligence agencies. According to an intelligence expert on the Cold War states will always seek for so-called ‘eyes on the ground’. Especially the United States of America is very reluctant to send their soldiers abroad, so one of the main criticisms against signature strikes by UAV’s, is that there is no more human confirmation involved (Scahill, 2016). However, according to this intelligence expert on the Cold War, it is not necessary at all to send one of your ‘own’. The former intelligence officer of the AIVD emphasizes the same thing:
20
“When you work abroad, you always cooperate with the intelligence agency of that country. If you need eyes on the ground, they can provide you with a local agent. You never send your own agent in.”
The combination of using both HUMINT and SIGINT is part of what agencies call ‘collection disciplines’. Ever since the Cold War, different INT’s are used in a single operation, meaning that an agency is never reliant on just HUMINT. Utilizing a single agent is too vulnerable and limited in scope. Sources are supposed to be complementary, and therefore agencies aim to combine human observations, satellite images, electronic signals, communications intelligence and all other sorts of INT’s acquired by technological means. This is not necessarily a new development, rather, it has always been a goal of many agencies to integrate these different kinds of INT’s. The intelligence expert on the Cold War illustrates this development with a quote on the search for Osama bin-Laden:
“With the use of satellites and drones the U.S. was able to get images of the alleged house of Osama bin-Laden. Occasionally they saw men moving around on the roof, but they were not able to identify bin-Laden as one of them. Therefore, they asked a local Pakistan doctor to make a tour through the neighbourhood for a cover-up vaccination campaign, in order to get a possible DNA sample of bin-Laden. This mission failed, however. He got captured and put into jail. But the point is: agencies need to use all different kinds of INT’s. And most of all: they need human confirmation.”
Examples often raised by the interviewees on the lack of human confirmation are the recent terror attacks in both Paris and Brussels. Questions they raise are, for example: What would have been prevented if there were agents in the district of Molenbeek in Brussels? Are intelligence agencies that focus solely on SIGINT not very naïve, since the Paris terrorists reportedly communicated through brief paper? It is difficult enough to trace down small groups or lone wolves, so how do we expect to find them by only using SIGINT methods?
Many news media reports refer to both these terror attacks as ‘intelligence failures’ and strengthen the rhetoric that intelligence agencies in their current capacities do not suffice (Calamur, 2016; Camilli, 2016; Robertson, 2015). According to the former intelligence officer of the AIVD, especially the lack of cooperation between certain agencies of different states is problematic and partly responsible for missing, for example, the proceedings of Salah Abdeslam (suspected terrorist of the Paris attacks). In his experience ever since the Cold War
21
the intelligence system of Belgium has proven to be insufficient on many different levels, and the recent terror attacks in Brussels and unawareness of Abdeslam’s whereabouts are just a mere example of this. Therefore, he recommends better cooperation between agencies, since some are better in getting HUMINT and others in SIGINT.
Having had ‘eyes on the ground’ in Brussels’ Molenbeek district would have made a great difference in identifying possible targets or terrorists, according to the same intelligence officer. He gets support from a terrorism expert who believes human infiltration in certain organizations is one of the keys to be successful. She illustrates this by raising the following questions:
“How can we be surprised that ISIS suddenly became so big? Did we not invest millions of dollars and euro’s in countries such as Iraq in the hope that we could create a more liberal and democratic state? Why were we not capable to trace that money to see where it ended up? Why did we not use local agents to infiltrate militias who used this money to set up their own state? Because let’s be honest: we do not acknowledge them as a state, but they certainly have the main characteristics.”
An intelligence expert and author of books on intelligence agencies and security services, agrees with this in a sense that he thinks it is ‘bizarre’ that the U.S. actually had to install a telecommunications network in Iraq after the invasion, in order to get their technical intelligence products:
“What about the people there who do not own phones? What about the local dialect of a certain tribe? Is anyone in the U.S. capable of translating this?”
This emphasis on the importance of human presence in intelligence gathering is meaningful for more than abovementioned reasons. HUMINT is not solely focused on infiltration, rather, a big part of the daily business of an intelligence agency is counterintelligence. According to the CIA (2008) the definition of this concept is as follows: “Counterintelligence embraces both ‘information gathered’ and ‘activities conducted’ to counter foreign intelligence threats”. Both forms of intelligence products mentioned in this definition can either be HUMINT or SIGINT. Until today, humans play a pivotal role in both practices.
The intelligence expert on the Cold War believes, for example, that very important information, such as characterizations of intelligence officers, agents or infiltrates, can only be
22
obtained by hearing a captured agent. Information such as the possible alcohol problem of a certain agent, or the amount of mistresses he visits in his spare time, can be of much value in certain operations. He believes that technical intelligence will never be able to find out things like this. A former historian of the AIVD, who was able to study the archives of the service and interview many (former) employees, states that the actual intentions of people can hardly be obtained by technical means. HUMINT is in their experience a more reliable and convenient method to ‘get inside the head’ of people.
Another example of this emphasis on human presence is provided by the intelligence expert on the Cold War. He refers to the “Stuxnet virus” the Americans and Israelis allegedly used as a cyber weapon to sabotage Iran’s nuclear program (Nakashima & Warrick, 2012):
“Both the American and Israelis were not capable of hacking the nuclear program from outside, so they eventually decided to put a virus on a flash drive and let one of the workers take it into the facility. Probably unknowing that it was contagious. The presence of a human in such operations is irreplaceable.”
Nevertheless, there are also a lot of disadvantages of using humans in operations. According to many interviewees debates about the reliability, effectiveness, high costs of training people, and durability of agents are the order of the day. For example, in line with Robertson’s (2015) observations, the former intelligence officer of the AIVD agrees that in HUMINT operations sometimes a disproportional amount of officers, sometimes up to twenty or thirty people, has to monitor a target in order not to lose sight of him or her.
The former historian of the AIVD emphasizes that especially with regard to reliability a lot of disadvantages can come up. She states how ‘agent loving’ can take place. That occurs when a ‘runner’ – an intermediate between an intelligence officer and the agent – starts to care too much for an agent. This runner can lose its independent stance and, for example, claim too much or infeasible compensations for its agent. Or, sometimes runners or intelligence officers need to manipulate their agents in order to get things done in an operation. Sometimes this can have a severe influence on their private life. She exemplified this with the troubles the BVD faced when they had to reincorporate agents who infiltrated the ‘Communistische Partij van Nederland’ (CPN) in the early ‘80’s of the 20th
century. It was very difficult to let these people return to their old lives without too much reputational damage; the aftercare was extensive and long lasting.
23
It is a widely supported notion amongst the interviewees that HUMINT and SIGINT cannot exist without each other, although both methods have their advantages and disadvantages. Therefore, even though the automation of observation processes in intelligence gathering - in this thesis referred to as SIGINT – is prevalent, the importance of human observation and information gathering has not declined. The uniqueness of the information provided after infiltration of an organization, or from the hearing of unmasked foreign agents, is timeless and co-exists next to technical information.
So, the narrative in media reports that the human presence in observation processes declines might, to a certain extent, be true, since more and more information comes through SIGINT methods. But, according to the former intelligence officer of the AIVD, the need for HUMINT will prevail and has actually not changed in size in the last decades. SIGINT methods and open-source intelligence (OSINT) have made it possible to acquire more data, and therefore it seems as if HUMINT practices are considered obsolete. But in practice the need for a human touch will always be present.
Furthermore, to a certain extent the importance of HUMINT alongside SIGINT is in line with Weizman’s concept of “forensic architecture” (2011: 103-104), since intelligence officers (or agents) are capable of interpreting ‘material evidence’ and therefore considered expert witnesses. On the other hand, the increase of SIGINT methods in ISR practices also means that in this “era of forensics” (2011: 115) governments and intelligence agencies are overcompensating the indeterminate and fragile voice of the victim. Even though Weizman refers to local witnesses of a certain operation in this case, it does imply that the presence of the ordinary human is less meaningful than it was before.
Therefore, this chapter shows how observation practices performed by humans are still at the light of day, and confirms Weizman’s notion that the problems associated with human testimony come back in current forensic practices. Thus, materials or technology do not speak for themselves; humans are necessary to create a rhetoric. A notion that is even more applicable to the next chapter, in which the analysis of data will be examined.
4.2 The uniqueness of a human touch in analysis practices
The second theme discussed here concerns the unique aspects of a human touch in analysis practices. After the interviews it appeared that human action is not just inevitable in observation processes, its importance is probably even more applicable to analysis practices. In this chapter the concept of analysis entails the interpretation of data by algorithms, the
24
interpretation of material remains by expert witnesses, and the presentation of an intelligence product by an intelligence officer after a process of observation and interpretation of data.
One of the main critiques in both the media and the academic world is the difficulty of coping with the incredible amount of data that SIGINT practices produce (Lyon, 2014; Read, Taithe & Mac Ginty; 2016). The former intelligence officer of the AIVD is one of those critics:
“By using satellites, SIGINT and all other kinds of technological intelligence methods we try to collect data. It will not take long before we try it by using glass fibre too. This way you gather so much data that you are not able to see the wood for the trees. Yes, algorithms are capable of discovering trends or patterns, but they cannot discover very specific threats. You will never find a guy like Salah Abdeslam by waiting for answers from an algorithm.”
This focus on the reliability of machines, in this case the objectivity of algorithms, is what concerns all the interviewees. It is extremely difficult to predict human behaviour, or label certain acts as ‘suspicious’. Because who decides what the concept of suspicious actually entails?
Therefore, consensus reigns amongst the interviewees about the importance of human impact in the process of analysis of data. First of all, for quite an obvious reason: the software responsible for interpretation is developed by humans, and that immediately raises questions. The terrorism expert is very explicit by stating the following:
“These algorithms are developed by people. Most of the time these are young engineers, including some who are capable of hacking. In my experience only a very select group of people is capable of understanding these algorithms. For example, I spoke to engineers from the ‘Belastingdienst’ [Tax and Customs Administration in the Netherlands] and they told me that their superiors have no clue about what they are doing. It was too difficult for them to comprehend and they – these engineers - did not have to bother trying to explain it.”
This is a clear example of how the software to analyse data is only understood by a small amount of people, whereas the tendency prevails among society and governmental agencies that it is a more reliant and trustworthy mechanism of interpreting information (Weizman, 2011). There is a paradox present in this development and therefore the importance of a human touch cannot be neglected. A PhD candidate on terrorism and security studies
25
elaborates on this paradox by stating that (meta-)data can only be interpreted in an effective way when human action is involved:
“Even when you have a large amount of meta-data and you let the algorithms do their job, it is still important to insert criteria somewhere. Every intelligence product is needed for a certain cause, or operation, imposed by national authorities. Humans are responsible for this, because it is a policy decision. Therefore you need criteria to establish a focused analysis. You need to create some sort of bottleneck. And after the first results you, as an analyst, need to reflect on that, sometimes with your management. So, in order to be effective, you need human actions.”
A former head of a Dutch intelligence agency agrees with the fact that the determination of search criteria is exclusively a human enterprise, and adds an interesting notion regarding the highly debated topic of privacy:
“I believe the government should be able to collect more data than it does at the moment. They need to save it somewhere, and only use it when they think someone poses a threat. Yes, a lot of people will fear for their privacy. But the work of an intelligence agency is to search for a needle in a haystack. Why are we so afraid of Big Data? Especially in the Netherlands, a country that in the last hundred years has showed that it has more than reliable governments. Why do people distrust them?”
According to all the interviewees, the public forgets about the infinite need of a human presence and action in ISR practices. Intelligence agencies already acknowledged this fact decades ago, and are now deploying their personnel more and more for both HUMINT and SIGINT purposes. They understand they act as servants of society. And in order to do so effectively, they understand that in analysis practices it is important to combine different methods of intelligence gathering. Gordon Corera (2016), a British author who publishes on computer technology and how intelligence agencies in the United Kingdom use this, explains how in the U.K. GCHQ and MI6 (Military Intelligence, Section 6, part of the Secret Intelligence Service (SIS)) are moving closer together in their operations. In his article (2016) he starts by underlining the infinite need for HUMINT, and how GCHQ can facilitate the MI6 officers (who try to recruit agents) with information, for example by identifying patterns of online activity. Nevertheless, he recognizes the balance is shifting; GCHQ is roughly double
26
the size of MI6: “Inside MI6, there's an understanding that they will need a new type of spy and everyone will need digital skills.”
This development is in contrast with the type of people that were deployed, for example, during the Cold War period in the Netherlands. The former intelligence officer from the AIVD and BVD recalls that almost exclusively ‘alpha’ oriented people were hired, because they had the right characteristics to train, deal with and influence possible agents. Also, these people were better qualified to stay in contact with agents in the field. Communication methods were, for example, letters written in codes (cryptography) and passed on via so-called ‘dead drops’: hidden places where someone dropped his or her information before it would eventually be picked up by a runner.
This old-fashioned exchange of data (intelligence) is still commonplace and, just as SIGINT information that can roll out of the printer at the office, this intelligence product needs to be presented. The former intelligence officer of the AIVD and BVD explains how this happened during his career (from 1970 until 2009):
“To be honest, nothing much has changed during that period. Yes, we developed more different methods of acquiring data. But as soon as I became team leader in my department, I had regular meetings with my analysts during which they had to present me their findings. It does not sound very exciting, but all I wanted was a piece of paper, or two, with bullet points and their major observations. I did not want to know their research methods or very specific data. I needed general findings which I could submit to my superiors: the political leaders.”
This shows how the actual analysis, the process of estimating what is important information in a certain operation, is always context-dependent. Political leaders are responsible for choosing what organization or government should be investigated. The responsibility of the intelligence agency is to come up with information that suits the demands of the political leader – in the case of the Netherlands the Minister of Interior and Kingdom Relations. According to the PhD candidate it is impossible to assign the decision of setting up criteria for investigation to a certain type of software. They are not flexible and cannot cope with nuances or morality:
“A critical assessment of sources is necessary. Therefore, I do not believe in the replacement of the human in this process. Yes, a computer is more accurate, able to cope with a bigger set of variables, and is better at predicting scenarios. But to actually interpret the findings and translate it into a specific research goal, that has to be done by a human.”
27
Sometimes, as in the case of the aforementioned Human Rights Watch report on drone strikes conducted by the Israelis, experts are enlisted to observe and analyse the material remains, in order to produce objective evidence of certain actions. The implications Weizman (2011) already acknowledges concerning this method of analysis, are underlined by most of the interviewees: are those so-called experts less human than others?
There is a common understanding amongst the interviewees that the analysis part in ISR practices should be the core competence of an intelligence agency, and for the relatively small AIVD in the Netherlands in particular. The PhD candidate describes this as follows:
“We must ensure that we do not get lost in an endless pursuit of Big Data collection. Since this will take the eye of the main goal of an intelligence agency: protecting our national security. The AIVD should carry out that they have a large brain trust and they have answers to many security related issues. Make the analysis part the core competence. It will give them more legitimacy, but on the other hand forces them to be more open and transparent.”
In a later chapter the interplay between the political leaders and intelligence agencies will be addressed, as well as the subsequent caveats regarding transparency and communication between the different parties. This chapter tried to show how the presence of the human has changed due to the advent of technological innovations in analysis practices, but simultaneously remained of great importance to intelligence agencies.
This change in physical activity of the human in data analysis specifically took place after the advent of Big Data. Collecting this much data is for a lot of intelligence agencies and governments an encouraging development in their fight against crime and terrorism, but how to effectively analyse them is a question that remains to be answered. The software, or algorithms, developed for this task are understood only by a small amount of people, whereas their superiors are the ones who actually have to make policy decisions on it. In that regard, the importance of a human touch in analysis of data has increased. As mentioned in particular by the PhD candidate, to effectively use these data, humans should be the ones to set up the criteria that will deliver a workable intelligence product. This is in line with the main findings of Read, Taithe & Mac Ginty (2016) who focus in their research on the utility of data if organizations are not able to cope with the large amount. Thus, the ideal of presenting a ‘real’ picture of a security setting by exclusively being reliant on SIGINT is ostensibly impossible (2016: 10-13).
