Author: India Bill Christopher Arputharaj is a Research Scholar affiliated with the Department of Strategic Technologies, School of National Security Studies, at the Central University of Gujarat, Vadodara Campus, Gujarat, India. His academic work is situated within the broader field of strategic technologies and national security studies. For all correspondence regarding this research, he can be reached at christopher240721001@cug.ac.in. Publisher: Fulcrum Analytics.
Backdrop:
Threat Landscape in the Sahel Region:
The
Sahelian theatre of operations, focusing on Burkina Faso, necessitates a
re-examination of Counter-Insurgency (COIN) and Counter-Terrorism (CT) tactics
in light of the dynamic nature of asymmetric warfare. The crisis traces back to
the 2011 downfall of Libyan state control, which had caused a flood of
sophisticated weapons and seasoned combatants into the Sahel region and hence
encouraged insurgencies in northern Mali as well as supporting organisations
like Jama’at Nusrat al-Islamwal-Muslimin (JNIM) and the Islamic State in the
Greater Sahara (ISGS).[i] By
2015, this instability had spilt over into Burkina Faso, in the form of
targeted violence in Ouagadougou to illustrate the vulnerabilities of the
state, stunt the economy and tourism in particular and damage national morale. The
conflict has escalated to a generalised rural insurgency, involving the
Burkinabe government and its supporters in a long-term fight in which
insurgents have control over the timing and tactics of engagements. This escalation
from a regional problem to a serious domestic insurgency highlights the
transnational and fluid threat these groups represent, regularly outmanoeuvring
traditional military action and exposing deep vulnerabilities in intelligence
and operational planning.[ii]
The insurgents take advantage of porous borders for prolonged operational
effectiveness, creating a cycle of tactical manoeuvres that do not produce
significant strategic dividends. Such a context underscores the imperative for
breakthrough technological solutions and an urgent re-examination of
operational models to reverse the adversary’s continued gains.[iii]
The
Operational Environment: Asymmetric Challenges in African COIN/CT:
The
following establishes the strategic challenges for traditional military forces
in Burkina Faso, which are compounded by an operational environment that aims
to degrade their performance. The war was presented as asymmetric warfare,
where insurgents used geographic advantages such as ungoverned territory and
open borders with Mali and Niger for conducting strategic operations. They
employed hybrid warfare strategies of conventional ambushes, attacks on fixed
military targets, and terrorist attacks on civilians, hindering counterattacks
by the military. These are compounded by scarce resources and intelligence
capabilities, which make it challenging for military forces to exercise control
and legitimacy as insurgent operations seek to strip the state's security
provision of public confidence.[iv] A
key issue is the intelligence shortfall that prevents decision-making and
constrains commanders from forestalling insurgent movements, resulting in
tactical surprise. The absence of integration within predictive intelligence
blocks end-to-end Pattern-of-Life (POL) analysis and the establishment of a
proven Common Operational Picture (COP), leading to enormous collateral damage
that is exploited by insurgents for propaganda. It compromises essential Human
Intelligence (HUMINT) in counterinsurgency operations, which leads to a
disconnect between governments and people that the insurgents exploit. Existing
reactive approaches, though providing temporary relief from short-term threats,
inadvertently support insurgent recruitment and amplify grievances. Overcoming
this paradox requires the creation of new capabilities aimed not only at
gathering data but at concentrating foresight to end the reactive engagement
cycle.[v]
Fig.
1: Heatmap Analysis of African Terror Incidents (1946 – 2024)
Data
Source: UCDP, Shapefile | QGIS 3.44.3, 2025 | Author
MAP
NOT TO SCALE | FOR ILLUSTRATION PURPOSES ONLY
AI-Driven
Predictive Analytics: Core Capabilities for the Modern Battlespace:
To
enhance operational clarity in intelligence, the employment of AI-driven
technologies is crucial, transitioning from descriptive to prescriptive
insights through advanced machine learning techniques that analyse diverse
datasets for hidden patterns.[vi] Utilising
spatio-temporal pattern analysis via deep learning models like TabNet (Tabular
Deep Neural Network Architecture), EBO (Effects-Based Operations) facilitates
the exploration of multi-dimensional data, including historical attack logs,
Signal Intelligence (SIGINT), Geospatial Intelligence (GEOINT), and Human
Intelligence (HUMINT) reports. This methodology unmasks non-linear
relationships, allowing for the forecast of likely hotspots and cyclical
patterns of attack, generating probabilistic terrain maps predicting insurgent
activity, thus upgrading intelligence from rudimentary reporting to proactive
predictive analysis. Additionally,
optimised ensemble classifiers, utilising Particle Swarm Optimisation (PSO),
Random Forest, and Xtreme Gradient Boosting (XGBoost) algorithms, improve
attack type classification accuracy, essential for determining possible
weaponry, targets, and tactics.[vii] The
unification of communication data in a Social Network Analysis (SNA) paradigm
enables the identification of attackers and leading actors in logistics,
finance, and propaganda, with High-Value Targets (HVTs) as the priority and
shifting operational tactics from attritional to targeted network disruption.
Deep Neural Networks (DNNs) and Gradient Boosting regression models provide
predictive logistics support needed for the anticipation of Mass Casualty
Incidents (MCI).[viii] These
models analyse anticipated attack behavior to provide estimations of likely
casualties, facilitate strategic preposition of medical equipment,
determination of aeromedical evacuation corridors, and the deployment of blood
supplies to minimise the overall impact of attacks. The Synthetic Minority
Over-Sampling Technique (SMOTE) also accounts for imbalances in data within
areas of conflict, emphasising important but infrequent events despite their
rarity over less effective frequent occurrences. Consequently, this integration
of technology converts raw data into usable battlefield intelligence, requiring
the creation of a new warfighting doctrine to implement effectively.[ix]
Fig
2: Taxonomy of AI Techniques in Counter-Terrorism
Source: Syllaidopoulos,
I., Ntalianis, K. S., & Salmon, I. (2025). A Comprehensive Survey on AI in
Counter-Terrorism and Cybersecurity: Challenges and Ethical Dimensions. IEEE
Access, 13, 91740–91764. https://doi.org/10.1109/ACCESS.2025.3572348
Applications
in COIN/CT Operations:
The
integration of Artificial Intelligence (AI) in the Concept of Operations
(CONOPS) of the military significantly enhances operational effectiveness by
improving the Observe-Orient-Decide-Act (OODA) loop.[x]
Leveraging an AI stack as part of the Intelligence Preparation of the
Operational Environment (IPOE), the system continually processes real-time and
historical information to generate a Dynamic Common Operational Picture (DCOP).[xi]
This DCOP facilitates nimble evaluations of the battlespace to provide
commanders with the ability to detect insurgent operations, forecast weapon
caches, and assess infiltration corridors, thus facilitating a focus on
particular missions rather than large-scale Intelligence, Surveillance, and
Reconnaissance operations. The
CONOPS also enables Precision Targeting and Effects-Based Operations by using
AI to detect High-Value Targets (HVTs) and key nodes, which enable the accurate
strikes that can disrupt insurgent supply chains, for instance, by targeting
logistical support staff rather than fighting many firefighters. It can also
assess non-kinetic measures like psychological operations and cyber actions,
thereby creating multi-domain challenges for adversaries.[xii]
Predictive Force Protection is improved through AI-generated threat assessments
that support proactive risk management for patrol routes and infrastructure,
leading to decreased casualties and enhanced operational efficiency. The system
also enhances Information Operations (IO) by analysing media sentiment and social
network trends, helping to counter misinformation and promote narratives that
strengthen state legitimacy.[xiii]
However, the deployment of this CONOPS introduces new digital vulnerabilities
and strategic risks, necessitating careful strategy formulation for effective
mapping and mitigation to maintain operational security.
Challenges,
Critical Limitations, and Strategic Risks:
In
AI-guided military operations, several weaknesses and strategic issues arise
that are similar to those associated with conventional warfare. Central to this
is the “garbage in, garbage out” phenomenon, which underscores data quality and
consolidation as crucial to successful predictive analytics.[xiv]
Some of the challenges include disparate reporting, biased historical data, and
poor standardised data-sharing among G5 Sahel partners, all of which compromise
algorithm training and escalate the threat of mission failures. Poor data
inputs can lead to mistaken tactical advice, potentially resulting in incidents
like friendly fire or civilian harm.[xv][xvi]
Algorithm bias is a significant concern; AI models trained on data from
specific ethnic areas, such as those populated by the Fulani, risk erroneously
associating that ethnicity with insurgency, thereby threatening operational
decisions and potentially harming community relations.[xvii] Moreover,
adversaries may employ tactics such as data poisoning and misinformation to
undermine predictive accuracy, creating a “fog of data” in combat situations,
which necessitates heightened vigilance regarding information superiority. The
lack of clear ethical and legal frameworks regarding the role of AI in deadly
actions makes compliance with International Humanitarian Law (IHL) more
difficult, especially regarding responsibility for AI-caused deaths.[xviii]
The current debate addresses whether the responsibility should rest with
operators, developers, or commanders. Ineffectiveness of Human-in-the-Loop
(HITL) systems and unclear Rules of Engagement (ROE) intensify ethical
dilemmas, jeopardising operational efficiency and legal legitimacy.[xix]
Future strategies must strive for careful equilibrium between embracing
technological innovation and maintaining rigorous ethical and technical
protections to prevent catastrophic outcomes.
Conclusion:
Military
tactics in Burkina Faso's war need to go beyond conventional counter-insurgency
by leveraging Artificial Intelligence (AI) for predictive analytics, from
reactive to proactive. A human-focused decision support system is needed for
accurate, timely results for complex operations. Having a Joint Inter-Agency AI
Fusion Cell (JIAFC) will focus on AI and machine learning endeavours, aided by
intelligence officers and data scientists. It focuses on Human-in-the-Loop
(HITL) procedures for human monitoring of all kinetic activities to maintain
adherence to International Humanitarian Law (IHL) and ethics. It is important
to address algorithmic bias with an inclusive Bias Testing and Mitigation
Program that comprises audits and methods such as the synthetic minority
over-sampling technique (SMOTE). Increased cooperation towards data standardisation
between ECOWAS and G5 Sahel countries is crucial for safe data sharing and
enhanced predictive powers, ultimately campaigning for responsible AI
integration in Counter-Insurgency and Counter-Terrorism (COIN/CT) operations.
[vi] Shao, Z., Qian, T.,
Sun, T., Wang, F., & Xu, Y. (2025). Spatial-temporal large models: A super
hub linking multiple scientific areas with artificial intelligence. The
Innovation, 6(2), 100763. https://doi.org/10.1016/j.xinn.2024.100763
[ix] Elreedy, D., Atiya, A.,
& Kamalov, F. (2023). A theoretical distribution analysis of synthetic
minority oversampling technique (SMOTE) for imbalanced learning. Machine
Learning, 113. https://doi.org/10.1007/s10994-022-06296-4
Comments
Post a Comment