Chapter I: Operational context of the FCAS

This first chapter examines the components of the FCAS (Future Air Combat
System), highlighting the complexity of this interconnected system that integrates
land, sea, air, space, and cyber environments. This chapter explores the technological
advances that constitute the FCAS and provides an in-depth understanding of the
operational framework in which the thesis evolves. This analysis redefines the roles
and responsibilities of the actors involved, particularly the pilots, and highlights the
importance of trust and information management in this system. This chapter is based
mainly on the FCAS presentations communicated by the Ministry of the Armed Forces
(Ministry of the Armed Forces, 2022a, 2022b, 2023) and an information report for the
Senate (Le Gleut & Conway-Mouret, 2020)

I.1. The components of the FCAS

The evolution of military threats and defense systems, driven by technological
advances and tactical adaptations, is redefining national defense strategies. Stealth
aircraft, cyber defense systems, drones, and hypervelocity missiles are at the heart of
future arsenals. To meet new battlefield challenges, defense strategy is evolving toward
increasing interconnectedness, integrating land, sea, air, space, and cyber
environments. This transition to Multi-Middle, Multi-City (M2MC) defense heralds the
development of the Future Combat Air System (FCAS), which embodies a paradigm
shift.

The FCAS is a European project of a set of interconnected air weapons

systems involving France, Germany, Spain and Belgium.

This project involves the development of a FCAS for each participating country,
which will connect existing national capabilities (Dassault Rafale for France,
Eurofighter Typhoon for Germany and Spain, with drones, surveillance aircraft, in-
flight refueling and command systems). The FCAS is structured around several
interdependent components, organized into two circlesconcentric (see Figure 1).

• At the center is the combat aircraft named NGF (Next Generation Fighter).
This aircraft performs functions such as interception, air-to-air defense,
and, in the French context, deterrence missions. The presence of a pilot

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
 on board is preferred for critical decisions where human judgment is
essential. However, the evolution of technologies and operational
doctrines could in the future redefine the role of humans within this
complex system.
• The first circle encompasses the NGF and the "Remote Carriers" as well as
the "combat cloud," together constituting the Next Generation Weapon
System (NGWS). "Remote Carriers" refer to unmanned systems, ranging
in size from a small drone to a larger vehicle. They perform defensive
saturation maneuvers, consisting of deploying swarms of aircraft to
overwhelm enemy defenses, as well as for decoy and intelligence-
gathering missions. These aircraft are also capable of guiding and
defending aircraft. They also benefit from a degree of autonomy enabled
by artificial intelligence. This autonomy makes it possible to manage
potential threats proactively and in coordination with combat aircraft.
These capabilities are all integrated into a "combat cloud," a collaborative
network designed to maximize the cohesion of defensive operations.
• The second circle, called FCAS, extends its scope to include existing national
capabilities, such as the Rafale and the Eurofighter. It also integrates
other elements, such as refueling aircraft, radar aircraft, satellites,
command, naval vessels, ground forces, and linked forces.

Figure 1: Representations of the FCAS components. In the center is the NGF,

the blue circle corresponds to the NGWS and the green circle constitutes the
FCAS. (Illustration modified from the FCAS project © DGA/MIP)

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
 In summary, the FCAS represents the development of a Next Generation Combat
Aircraft (NGF), combined with drones (Remote Carriers). This combination is
interconnected within a combat cloud. The NGWS (NGF + Remote Carriers, within a
combat cloud) is capable of operating either autonomously or in a network with air,
naval, land, or space-based combat or command systems (FCAS). Within the FCAS,
each element—whether the NGF aircraft, the Remote Carriers, or the various command
and control modules—is designed to collaborate within the combat cloud.

I.2. The combat cloud

The Combat Cloud is a collaborative network designed to integrate and
synchronize data and communications within the armed forces.

By combining data processing and storage capabilities with artificial intelligence

and robust cybersecurity measures, this platform centralizes information from various
environments, such as land, sea, air, and space. This centralization of information
allows for management and effective dissemination of strategic intelligence, thus
ensuring air superiority and facilitating rapid decisions on the battlefield.

However, this technological evolution is not without risks; the main danger
comes from cyberelectronic attacks, a threat that results from the convergence of
electronic warfare and cyber warfare. There are two types of cyberelectronic threats:
on the one hand, jamming affecting the transmission network and sensors, and on the
other hand, attacks related to Offensive Cyber Warfare (OCW), which can target the
entire cloud (Gros, 2019).

Among these threats, Offensive Cyber Warfare (OCW) appears to be of particular

concern. Indeed, reports from the US Director of Operational Test and Evaluation
(DOT&E, 2021, 2022, 2023) highlight the "cyber vulnerabilities" of many US systems,
including recent systems that are supposed to integrate measures against this threat,
such as the F-35. These reports highlight the extent of cyber threats that continue to
pose a challenge to even the most sophisticated systems.

Furthermore, the increase in interconnections between systems amplifies the

possibilities of electronic intrusions in the cloud, thus raising the risks of systemic
effects. In other words, a poorly secured combat cloud could, faced with a competent
adversary, risk partial or complete paralysis of the entire force. The combat cloud must
therefore operate in potentially hostile environments where cyber threats are

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
omnipresent. As a result, the collaborative network requires robustness against
interference and attacks to preserve the integrity and confidentiality of transmitted
data.

Data management is equally essential for the effective operation of the FCAS.
The data collected by the stakeholders involved must be carefully sorted, processed,
and analyzed to turn it into actionable intelligence for operators. Currently, the Rafale
is already networked, but the pilot uses mainly his own sensors and, to a lesser extent,
information provided by the network. The FCAS will be characterized by developments
such as improved sensor capabilities, optimized use of the electromagnetic spectrum,
increased storage capacity, and a cloud of combat gathering a large amount of
information. (Le Gleut & Conway-Mouret, 2020).

Given these developments, access to a much larger volume of data will be

possible. Data management and processing will be fully automated via the network,
independent of the pilot, who will only have access to merged and synthesized data.
The pilot's role will thus evolve into that of a supervisor of automated processes,
supported by a "virtual assistant" based on artificial intelligence (AI), (Ministry of the
Armed Forces, 2023). This AI-based virtual assistant will act as a strategic advisor,
reducing the pilot's cognitive load and minimizing the risk of information overload.
This will allow the pilot to focus on critical decisions.

I.3. Decision-making within the FCAS

In this context, decision-making within the FCAS becomes particularly complex,
involving the synthesis and interpretation of data from multiple sources. This
complexity is amplified by the need to quickly process this information to effectively
respond to the demands of the modern battlefield. The pilot, at the center of this
system, must navigate between information from human operators, analyses provided
by the virtual assistant, and feedback from basic automations (see Figure 2).

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
 Figure 2: Diagram of the different sources of information available to
the pilot within the framework of the FCAS.

• Human contributions, including tactical communications and chain-

of command commands, provide often urgent context.
• Various AI-based systems, such as virtual agents, enhance the
decision-making process by analyzing massive volumes of data at a
speed and accuracy unattainable by humans alone. These systems can
anticipate threat scenarios, suggest tactical maneuvers, or optimize
flight paths based on changing conditions.
• Data from basic automation, such as sensor readings and navigation
systems, provides vital technical information.

Managing this diverse information from different sources in a very short time
increases the pilot's cognitive load and underlines the importance of an advanced
decision support system such as a "virtual assistant". By reducing the pilot's cognitive
load, this "virtual assistant" allows the pilot to focus on critical tactical decisions, thus
enabling mission success.

I.4. The example of the CAS assignment

The example of a close air support mission concretely illustrates this operational

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
complexity. The key actors in this mission include an aircraft and a Joint Terminal
Attack Controller (JTAC) who are integrated with the ground unit to request support
and coordinate or guide a strike. (CICDE, 2014). To illustrate this operational
complexity concretely, consider the example of a close air support (CAS) mission. The
key actors in this mission include the effector aircraft and the Joint Terminal Attack
Controller (JTAC) who is integrated with the ground unit to request support and
coordinate or guide a strike. The mission sequence is as follows: following a request
from its command or based on its own observations, the JTAC requests an air strike.
Consequently, the aircraft establishes radio contact with the JTAC, who then provides
a formatted briefing, the "9 Line brief." This briefing details several pieces of
information: the heading to be taken by the aircraft, the distance to the target, the
altitude, the description and coordinates of the target, the presence of friendly forces
in the area, the type of marking to be carried out by the JTAC, as well as additional
notes on surface-to-air threats, necessary coordination measures (such as
simultaneous artillery fire), the type of munitions required and the specific approach
to be executed for the strike. After receiving the briefing, the pilot and the JTAC
synchronize their awareness of the situation and confirm the location of the target. The
pilot then prepares his approach, and once in position, the JTAC gives him the final
authorization, or "clearance », to proceed with typing.

However, communications rely primarily on radio, and the exchanges between

the JTAC and the pilot can sometimes last for several minutes, to ensure that the target
hit is the correct one and to avoid collateral damage. Despite a standardized NATO
procedure, this synchronization can sometimes be subject to language errors and
interpretation.

This is why Close Air Support (CAS) is now assisted by digital technologies, such
as tactical data links (TDLs), to reduce these risks of misunderstanding and speed up
the decision-making process. This type of mission is called Digitally Aided Close Air
Support (DACAS). Although radio remains essential for fire authorizations or mission
cancellations, the JTAC can now transmit its requests via the Variable Message Format
(VMF), an LDT format chosen by ground forces because it is compatible with
conventional radio equipment. The details of the "9 Line brief" are transmitted either
by VMF message if the aircraft are equipped, or in several messages via the L16. The
L16 is a tactical data link for the exchange of tactical information between military
units. In addition to the JTAC brief, the pilot can prepare his intervention by consulting
the position of friendly forces, via the L16 and the Blue Force Tracking server, which
collects the precise positions of ground forces. Once contact is established, scanning
facilitates the transmission of targeted images from the aircraft's nacelle, allowing for
precise target identification. Once these images are received, the JTAC can confirm the

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
target.

DACAS, however, faces several challenges, such as differences in security levels

between the JTAC and the aircraft, the latter operating on a classified L16 link. This
situation prevents the pilot from directly integrating target data into his Navigation
and Attack System (ANS), forcing him to use a separate tool. In addition, complexity
increases with the difficulty of correlating data extracted from servers with that
provided by the JTAC during the terminal phase of the mission, thus increasing the
risk of errors that could compromise decision-making.

With the introduction of a combat cloud, the close air support (CAS) mission
could be transformed by increasing the speed and richness of information sharing, as
well as by better data exploitation by each stakeholder. In this new context, the JTAC
could directly indicate information relating to a strike on the combat cloud, making it
instantly accessible to the pilot. The NGF navigation system would then integrate not
only the target details (similar to DACAS), but also information from remote carriers
on reconnaissance missions. An onboard artificial agent would fuse all this data to
propose tactical options or to indicate potential data inconsistencies.

However, this procedure also carries risks, particularly related to the

cybersecurity of the combat cloud. There is a risk that combat cloud data could be
compromised due to a cybersecurity threat, thus affecting the reliability of the data.
However, the pilot, as well as the other operators involved in the mission, must be able
to trust the data transmitted by various partners, which may include other military
units, automated devices, satellites, remote carriers, or allied systems. Faced with this
situation, the virtual assistant can alert the pilot of inconsistencies detected in the
system. This function allows the pilot to make more reliable decisions, with a clear
understanding of the limitations and reliability of the available information. This
interdependence between technology and human interactions creates a framework
where trust in the accuracy of shared information is essential in the battlefield of the
future. The pilot must be able to operate with the certainty that the information on
which they base their decisions is not only up-to-date, but also reliable. This trust in
the veracity of the data is vital to maintaining the integrity of the operation and the
safety of the personnel involved.

The diversity of information sources—whether human, automated, or AI—adds

an additional layer of complexity to the perception of information reliability. Pilots
perceive and assess reliability differently depending on the nature of the source sending
them the information (Madhavan & Wiegmann, 2007a). For example, data from
human sources may be perceived as more subjective, yet rich in context and nuance,

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
than information from automated systems. This is because human information
incorporates personal judgments, experiences, and interpretations that may offer a
biased view of the situation. Automated information, from sensors or monitoring
systems, are often seen as more objective and technically reliable. These systems are
designed to collect and analyze data consistently and accurately, minimizing the
influence of human error. However, they can lack contextualization and be prone to
systemic errors or algorithmic bias, limiting their ability to adapt to unforeseen
situations or interpret data in nuanced ways.

Finally, AI-powered data, which analyzes and synthesizes large data sets, is
valued for its ability to quickly process complex information. AI systems can identify
patterns and trends that might escape traditional human or automated analysis.
However, they can also raise concerns about their transparency and interpretability,
particularly when they are complex "black box" algorithms whose decisions are not
always easily understood by human operators.

I.5. Problem and reading guide

Integrating these various sources of information into the pilot's decision-making
process requires a complex assessment of the reliability of each piece of information
depending on the nature of the source. This assessment of the reliability of information
is essential for real-time decision-making on the battlefield, particularly in the context
of the FCAS where the pilot's main tasks include supervision and strategic decision-
making. Trust in the reliability of information depends on the nature of the source. This
interdependence raises a fundamental issue for the FCAS:

How, in a complex sociotechnical environment, does trust in different

partners influence the pilot's decision-making?

The issue raised leads to the need to understand the different components of trust based
on the various types of sources present within the FCAS. Indeed, each source – whether
human, automated, or AI-derived – has distinct characteristics that influence how trust
is granted and maintained.

Therefore, Chapter 2 explores the concept of trust by providing a state-of-the-art

review of interpersonal trust and trust in automation. The state of the art establishes
the theoretical foundations of this thesis and proposes a definition of trust. Chapter 3
focuses on the dangers associated with the poor calibration of trust. In relation to the
problem and the operational framework, this third chapter explores the problems of

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
overtrust in decision-making tasks, called automation bias. This bias is detailed and
challenged by several empirical studies. This challenge allows the emergence of a new
paradigm: that of the dual advisor. This paradigm is defined and explored in depth in
the operational context. However, very few studies have focused on this paradigm in
conflict situations. Chapter 4 proposes a model of trust within the framework of the
dual advisor paradigm, allowing for the anticipation of decision-making and, by
extension, the analysis of the automation bias that can result from it. This model is
based on the theoretical foundations developed in Chapters 2 and 3. Together,
Chapters 2, 3 and 4 constitute the theoretical part of this thesis, providing the necessary
foundations for an analysis of trust mechanisms in the complex context of FCAS.

Chapter 5 presents the experimental methodology developed to empirically test

the hypotheses derived from the previously developed model. It describes the design
of the experimental campaign, the variables studied, as well as the measures and
procedures used to analyze decision-making within the framework of the double
advisor paradigm. It also outlines the contextual differences and methodological
choices of the experiment. Chapter 6 presents the pretests carried out to construct the
experimental material. Chapters 7 to 11 are devoted to the detailed presentation of the
experiments conducted within the framework of the experimental campaign. Each
chapter presents the results, analyses, as well as discussions specific to each
experiment. Chapter 12 is dedicated to a cross-sectional comparative analysis of the
different experiments presented in the previous chapters. Thus, chapters 5 to 12 form
the entire experimental part of this thesis.

Chapter 13 discusses the results of the experimental campaign by comparing

them with the hypotheses of the developed model. Chapter 14 addresses the limitations
of the study, particularly in terms of ecological validity and generalizability of the
results. It proposes perspectives for future research, such as the empirical
improvement of the model and the introduction of realistic contexts. Chapter 15
proposes recommendations that are structured around two main axes: on the one
hand, user training to optimize their understanding and use of the

AI systems, and on the other hand, the design of these systems so that they are
better adapted to the operational needs of users. The objective is to ensure harmonious
and effective collaboration between humans and the system, while strengthening trust
and acceptance of automated technologies in the context of FCAS.

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
Bibliography
CICDE. (2014). Allied Joint Doctrine for Close Air Support and Air Interdiction.
[Link]

Gros, P. (2019). Le “cloud tactique”, un élément essentiel du système de combat aérien

futur. Fondation pour la Recherche Stratégique.
[Link]
systeme-combat-aerien-futur-2019

DOT&E (Director of Operational Test and Evaluation). (2021, 2022, 2023). Annual Reports
on US Defense Systems (inclus les vulnérabilités du F-35)

Le Gleut, R., & Conway-Mouret, H. (2020). Rapport d’information n° 642 (2019-2020).

[Link]

Ministère des Armées. (2022a). FCAS : la France se félicite de l’accord.

[Link]

Ministère des Armées. (2022b). FCAS : notification du contrat pour la prochaine phase du
projet. [Link]
phase-du-projet

Ministère des Armées. (2023). « Le FCAS est bien plus qu’un avion ».
[Link]
avion

Madhavan, P., & Wiegmann, D. A. (2007a). Effects of information source, pedigree, and
reliability on operator interaction with decision support systems. Human Factors, 49(5),
773–785. [Link]

Madhavan, P., & Wiegmann, D. A. (2007b). Similarities and differences between human–
human and human–automation trust: An integrative review. Theoretical Issues in
Ergonomics Science, 8(4), 277–301. [Link]

Madhavan, P., Wiegmann, D. A., & Lacson, F. C. (2006). Automation failures on tasks
easily performed by operators undermine trust in automated aids. Human Factors, 48(2),
241–256. [Link]

Hoff, K. A., & Bashir, M. (2015). Trust in automation: Integrating empirical evidence on
factors that influence trust. Human Factors, 57(3), 407–434.
[Link]

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.
Kaplan, A. D., Kessler, T. T., Brill, J. C., & Hancock, P. A. (2021). Trust in Artificial
Intelligence: Meta-Analytic Findings. Human Factors, 65(2), 337–359.
[Link]

Schaefer, K. E., Chen, J. Y. C., Szalma, J. L., & Hancock, P. A. (2016). A Meta-Analysis of
Factors Influencing the Development of Trust in Automation. Human Factors, 58(3), 377–
400. [Link]

This work is a translated and adapted version of Chapter 1 of my doctoral dissertation defended in 2024 at the École de l’Air et
de l’Espace, France.