**6.2 Orient: a better analysis**

Remotely seeing will make it possible to identify a potential target from afar, to discriminate it (is the target a combatant) and to characterise its behaviour (is it hostile or not). If these criteria are met, the target becomes a potential target that can easily be geolocated, this information will then be transmitted to the

**Figure 1.** *OODA cycle time reduction: A better reactivity.*

decision-making levels. The gain here is that of anticipating the analysis for better decision-making.

The leader will also be able to rely on the automatic processing of data acquired within the digital environment of the battlefield. Faced with the potential 'infobesity' of the battlefield, artificial intelligence will enable massive data processing, subject to the availability of a computing capacity directly embedded in remote robotic platforms, or by remote processing of information via long-distance communications. It will allow constant monitoring of the analysis of captured images or sounds, a task that the best human experts can only supervise because they are subject to fatigue and inattention. This is particularly the case with satellite images or images captured by surveillance drones, which can monitor an area 24 hours a day. Finally, it will also enable the detection of weak signals that would be invisible to humans, by correlation between several distinct events, or by cross-checking.

There are still two essential components to the analysis of the situation that a machine can never integrate. Firstly, instinct and intuition, which a machine cannot have and which are the fruit of a life-long learning of human experience, and secondly, the transcendence of military action which only a metaphysical dimension in the literal sense can provide.

#### **6.3 Decide: a better reaction**

The military commander is the decision-maker for military action. It is therefore up to him to take the decision according to the information at his disposal. He can of course rely on a deputy or on operational advisers who help him analyse the situation, if time permits.

For example, France is intervening in Mali and the Sahara as part of the *Barkhane* military operation to combat Salafist jihadist armed groups infiltrating the entire Sahel region. Launched on 1 August 2014, this operation replaces operations *Serval* and *Épervier*. The following scenario is fictitious: an armed Reaper drone of the French army flies over a region of the Malian desert at night and its cameras (incorporating AI for automatic motion detection processing of the captured images) detect a suspicious movement. The sensor operator of the drone is alerted and zooms in on the area to detect a jihadist 4x4 occupied by armed personnel via its Infrared camera. This vehicle is moving towards a village 20 kilometres away. Setting up an operation with Special Forces is not possible because they are not in the area, and there is a great risk that the occupants of the 4x4 will disperse once they reach the village. The legal advisor on duty quickly confirms the possibility of the drones firing on the target because no collateral damage is possible in this desert area. The head of the operation decides to give the order to fire the drone.

This example clearly shows the drastic reduction in the OODA decision cycle offered by the new technologies: the chief detects and is informed as soon as possible by an automatic detection of a suspicious movement of an enemy vehicle. He confirms with his image operator the Positive identification (PID) of the target as an enemy. He then reports it to his hierarchy and receives the order to open fire. He can thus, in compliance with IHL, open fire from a distance. The enemy has not even spotted him.

There are still situations where time is critical and the leader will not have time to make a decision due to the rapidity of the attack. The automation of response processes then becomes a possible option, i.e. he can delegate to a machine the possibility of giving an appropriate response to a situation by itself. This is already the case with missiles or ballistic threats, which require armies to use automatic systems to counter them. This requires automatic systems that are faster and more precise than human beings (e.g. coupling weapons and radar). Tomorrow, faced with future systems

that will develop unpredictable trajectory strategies (enemy missiles with AI), faced with saturating threats that risk overwhelming our defences, faced with swarms of offensive robots, our systems will have to adapt in real time to counter the threat. Only a certain autonomy of the defensive systems will make it possible to face them, an autonomy which will have to remain under the control of the leader having these systems at his disposal.

#### **6.4 Act: a quicker and more accurate reaction**

A quicker reaction: A man reacts in a few seconds, the machine in a few milliseconds or less. Where a human thinks in a few seconds for the best, the machine will analyse parameters in a few milliseconds and propose a response in near real time.

A more accurate action: A human shooter who moves, breathes and shakes is less accurate than a machine that does not move, breathe or shake because it is not subject to emotion. Precision in action will therefore increasingly be the prerogative of the machine.

The outcome of a fight or a counter-measure may depend on these factors 10 or 100 seconds to a thousand seconds.
