AI military weapons should be ban

profilejuankverbel
E3.9.pdf

MT 11/2019 - 5

Marco Giulio Barone

Artificial Intelligence (AI) and military hierarchy

A rtificial Intelligence (AI) is amongst trending topics in today’s military procurement. Plenty of systems of weapons feature or will feature AI elements. AI is reshaping C4ISR functions as well, as related applications will allow for unprecedented real-time com­ prehensive situational awareness. So, in a conflict between peers or near-peers, tempo will be the key to victory more than assets. This is only part of the technology revolution that will allow armed forces to win wars through superior information rather than through superior fire­ power (something difficult to accept in military spheres). Another part is about how this becomes possible. Technological advantage is limited if not coupled with effective employment strategies.

Marco Giulio Barone is a political-m ilitary analyst based in

Paris, and a regular contributor to Monch magazines.

In other words, the faster situational awareness is distributed to blue forces on the battlefield and exploited for taking active decisions, the more victory will be likely. Pyramidal hierarchy-based decision-making processes require time to pass from strategic decisions to the lowest echelons on the ground. Hence, they are not the most proficient way of exploiting information technology advantage. One of the most significant lessons learned from Cold War’s local conflicts (e.g. Vietnam, Arab-lsraeli conflicts, etc.) was that doctrines based on rigid levels of authority - with high levels of command dictating attack patterns, philosophy of engagement, weapon firing, and communi­ cation procedures to the lowest echelons - were less proficient than mixed approaches of top-down and bottom-up levels of authority. In addition, decapitating the top of the pyramid during operations would be enough for winning the battle (or the war! Future conventional wars would unlikely last for years). Therefore, the need for responsiveness and resilience at the same time have found a solution: decision by nodes. There is already a consistent operational literature on it. High value assets can be replaced by networks of smaller assets, all contributing to the big picture or to larger operations. For instance, this is a key enabler delivered by 5th generation fighters that can serve as communication nodes, can perform electronic warfare (EW) duties, or geolocate targets for other assets. Crowdsourced information is also the formula used by satellite constellations of small satellites. AI applications promise to allow for extensive use of the same princi­ ples in all operational domains (air, land, sea, space, and cyberspace) thanks to Al-based applications’ capability of gathering, analysing, and disseminating relevant information to large networks. In addition, AI capabilities that will be fielded can allow the force application design­ ers to devolve more authority to the lowest echelon, even to the single man if needed (for instance, a fighter pilot). Technology challenges are many, such as the need of feeding AI applications with huge and con­ stant amounts of data or the increase of energy efficiency - AI applica­ tions are electric energy consuming. However, these are not existential for the future of AI based application for defence. The most unknown and underestimated obstacle to the success of AI on the battlefield is military officials’ mind set. During conferences and defence exhibitions, plenty of military officials praise industry advancements in the field, contribute to company presentations, and highlight the potential of AI applications to boost situational awareness on the battlefield. However, when questions arise about the revolution it brings to decision-making and changes expected in military levels of authority on the battlefield, answers are much less enthusiastic. Most military officials appreciate the possibility of gaining the full control of the battlefield through information superiority as well as of possessing responsive automatic assets immediately reacting to

orders on the tactic table. Yet, when they are questioned on the fact that Al-powered applications will bebattlefield-wide and will allow the lowest echelons to take auto­ nomous informed decisions, they start looking diffident. U.S., French, British, Italian and other countries’ officials share the fear that their “ personal power” in deciding the outcome of a battle is under attack. Their unconscious reflex is to hold position and resist to whatever might question their role of commanders. When industry officials or journalists underline that a major reflection on levels of authority, applied tactics, and flexibility of leadership is needed to better take advantage of AI appli­ cations, the answer is unanimously (and incredibly!) the same: “We can’t do everything at a time. We have our traditions and habits, and the military has worked well like it is. When the new generation of officials now entering academies will be at our place, they will be more ad­ equate than us to embrace change” . The logic behind the statement seems to make sense. But tempo will be vital in this field as well. A young lieutenant needs decades to become colonel or general. AI evolutions take months or years to advance, and the world evolves at so fast pace that one potentially disruptive technology arises on a weekly basis. Should the military fail to kick an internal reflection on this topic, procuring expensive cutting-edge AI appli­ cations will not win battle per se. Hampering or limiting AI develop­ ments not to trigger such a debate would be helpless as well.

Copyright of Military Technology is the property of Monch Publishing Group and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.