Is it a paradoxical injunction to ask the Forces to innovate? If "only victory counts," why not use even old means if they bring victory again?! Everything must change so that nothing changes, Lampedusa said in The Cheetah. Hence the question: what would be the right use of innovation within the Ministry of the Armed Forces?

For my part, I see three areas of indispensable and permanent innovation:

• The mindset

• The weapon system

• Artificial intelligence

The state of mind, or mindset , is by far the most complex. By definition, an officer is a man or woman of action. After a phase of observation and reflection comes the taking of action, which cannot leave room for doubt, in the full exercise of responsibility and trust. The act of taking action itself generates new results, information that must be collected and analysed in order to guide new actions that are as relevant as possible. The officer must learn to master these two very different ways of making his brain work. Otherwise, he would just be a blind bull. Knowing how to go from methodical doubt to conviction and vice versa is not the most common prerogative among humans. Festinger, in 1957, demonstrated this with his theory of cognitive dissonance. In short, we never waver, and reconstruct history to prove ourselves right. Kahneman, winner of the Nobel Prize in Economics, describes in System 1/System 2 dozens of absurd decisions related to the imperfect functioning of our brains. Yet flexibility of mind can be learned, and must be nurtured. Having read Marc Bloch's The Strange Defeat, I include myself among the civilians who must participate and get involved in the knowledge of doctrine, and create a continuum between the political and the military elites. "Peace is too serious a thing to be entrusted to pacifists," one might say, parodying Georges Clemenceau. This state of mind will apply continuously to (extract from the seminar of last 12 October of the Centre de Doctrine d'Emploi des Forces):

• maintaining operational superiority over an innovative enemy,

• to improve the soldier's daily life in the neighbourhood and on operations,

• develop opportunity innovation and ensure consistency with planned time,

• to be able to accelerate the realization of projects by removing technical obstacles (financing / purchase, employment authorization...) organizational and cultural,

• ensure decision-making and financial subsidiarity.

The second field of per- manent innovation exercises is the weaponssystem .Rather than imagining it in the hands of the DGA and the major industrialists involved, more frequent exchanges between future users and prototypes at different stages of their development are to be recommended, albeit in the expert hands of the DGA and the major industrialists. Integrating other players such as academics and start-ups would certainly be a factor of creativity. The so-called "agile" mode of development in civil industry also presents multiple interests within the Forces. Thus, the development of nanodrone swarms is likely to lead to the revision of a large number of tactical schemes, inherited from the past, when the trajectory of bullets was still straight. I should add that these nanodrones could discharge not lethal charges, but poisons with more or less deleterious effects, which would make it possible to begin negotiations, with the antidote in hand. These nanodrones could be amphibious, crawling, flying, stealthy, and even small enough to enter the enemy's body (nanorobots are used in surgery to unclog coronary arteries). Imagination must be limitless, intellectual and organizational. In other words: what would we dream about that would make our enemies have nightmares?

The third area of permanent innovation that needs to be opened up urgently is that ofartificial intelligence.Two pieces of information only need to be known in order to draw immediate lessons from it:

• Watson, IBM's artificial intelligence, has been fed 4 billion medical images, with the reports of operations that followed; its reliability rate is already 85%, compared to a world average of 65% for radiologists.

• Alpha GO Zero is Google's latest generation artificial intelligence, the so-called "self-learning" branch. After four hours of training, simulating billions of games against itself, after reading the game rules page, it beat 100 games at 0 Alpha GO - the artificial intelligence that had beaten the world go champion after having ingested billions of games... played by men. In other words, a self-learning AI can play perfectly, instead of just playing better than men.

Watson's CTO told us that IBM's AI, fueled by thousands of air battles, has beaten every US Air Force pilot in the simulator. The General who commands these pilots said "I've never seen such a vicious intelligence". And Putin said: "Whoever has AI will own the world".

The development of the AI at MinDef will have to be based on multiple experiments, by the users themselves, without the aristocracy of capability scorning the innovation of opportunity in short loops. Achieving the convergence of operational and algorithmic forces, "from data to decision", will involve an inevitable succession of trial and error. DARPA, the US Army's R&D centre with an annual budget of 3 billion dollars, seeks to ensure that more than 70% of its projects fail. Otherwise, according to them, it would mean that they were not ambitious enough.

France has one of the best mathematics schools in the world, as well as brilliant engineers. The qualities of its officers, who have dual experience in operations and central administration, are also recognized far beyond our borders. We have all the ingredients to ensure our sovereignty, provided that our ability to challenge ourselves is permanent. Henri Atlan summed it up very well in his book Entre le cristal et la fumée (Between crystal and smoke): to be alive is to know how to stand between the smoke, which looks like nothing, but which resists any shock, and crystal, a marvel of organization and beauty, which the slightest noise can dislocate. Let us know how to put order where and when it is needed, but also a little chaos, sometimes, "to give birth to a dancing star" (Nietzsche).

Innovation is a successful invention or idea. General Dempsey, former US Chief of the Joint Chiefs of Staff, placed it in the three qualities of a leader, with curiosity active curiosity and adaptability, "in order to anticipate changes in the world and its environment, with the team for which one is responsible". Rarely has the world changed so quickly, so strongly, so brutally.