Armed forces are turning to generic technologies, like AI 

Eight months since the begin­ning of the con­flict in Ukraine, many les­sons have been learned from this con­flict, not­ably the return to high intens­ity. One of the major issues arising from this, which goes bey­ond the mil­it­ary field, is the ques­tion of integ­rat­ing new tech­no­lo­gies to provide an oper­a­tion­al advant­age for the armed forces. 

Part­ner­ships, sup­por­ted by the US Depart­ment of Defense, with Microsoft or Amazon Web Ser­vices (AWS) to host Ukrain­i­an data that could be sens­it­ive to Rus­si­an cyber-attacks, or the use of Elon Musk’s Starlink mobile ter­min­als to con­nect to the Inter­net via satel­lite, have high­lighted the import­ance of the rap­id adop­tion of gen­er­al-pur­pose tech­no­lo­gies for mil­it­ary needs, and their effect­ive­ness on the battlefield.

In driv­ing the reform of its defence innov­a­tion eco­sys­tem, the French Min­istry of Defence has, in recent years, taken on a more entre­pren­eur­i­al and stra­tegic role, with a more exper­i­ment­al approach to innov­a­tion policy¹. Nev­er­the­less, the ques­tion remains as to the capa­city of our defence industry and our armed forces to be able to rap­idly adopt all these gen­er­al-pur­pose tech­no­lo­gies (GPTs) – such as arti­fi­cial intel­li­gence or quantum tech­no­lo­gies, at very dif­fer­ent levels of development.

Gen­er­al-pur­pose tech­no­lo­gies can be defined as those cap­able of con­tinu­ous future improve­ments and with the poten­tial to serve as the basis for com­ple­ment­ary innov­a­tions in related applic­a­tion areas (Teece, 2018). As such, a gen­er­al-pur­pose tech­no­logy has such an impact on over­all pro­ductiv­ity growth that it becomes ubi­quit­ous, as was the case with the inter­net in the 2000s.

With more and more prac­tic­al applic­a­tions in all sec­tors fol­low­ing sig­ni­fic­ant work in basic research, arti­fi­cial intel­li­gence (AI) has the poten­tial to become a gen­er­al-pur­pose tech­no­logy and thus trans­form the eco­nomy, nation­al secur­ity, and soci­ety as a whole.

Armed forces around the world are increas­ingly look­ing to AI as a key tech­no­logy for their long-term strategies and plan­ning. With tech giants such as Google and Amazon at the fore­front of AI innov­a­tion, major US defence com­pan­ies are under pres­sure to step up their AI activ­it­ies to keep pace with the com­mer­cial market.

100 mil­lion a year in arti­fi­cial intel­li­gence for defence, this rep­res­ents only 0.6% of the French armed forces’ equip­ment budget, where­as the Amer­ic­ans invest four times as much for an equi­val­ent budget².

Bey­ond invest­ment, sev­er­al bar­ri­ers to the adop­tion of these gen­er­al-pur­pose tech­no­lo­gies per­sist. From a tech­nic­al point of view, the vul­ner­ab­il­ity of the sup­ply of real data to train the algorithms (Osoba and Welser, 2017) is often the cause of unsat­is­fact­ory per­form­ance. How­ever, the prob­lem of the con­fid­en­ti­al­ity of mil­it­ary data has often, at least in dis­course, been an argu­ment that can hide anoth­er prob­lem: that of the gov­ernance of data, which is too often left to the indus­tri­al own­ers of plat­forms and weapon systems.

AI can have a crit­ic­al role in oper­a­tion­al plan­ning and decision-making.

In terms of user aware­ness, there are still oth­er obstacles, often linked to the unequal know­ledge of decision-makers of the uses of data and their added value for mil­it­ary oper­a­tions. This can be explained by the fact that sys­tems integ­rat­ing AI do not all enter the tech­nic­al rep­er­toire of the armed forces at the same time and with the same effi­ciency, but also by cul­tur­al and organ­isa­tion­al bar­ri­ers between the defence sec­tor and the civil­ian sec­tor. A strong dis­in­cent­ive to take risks, as well as pro­cure­ment pro­cesses that are not adap­ted to this type of tech­no­logy, are often enough to dis­cour­age any ini­ti­at­ive. Ulti­mately, accord­ing to the fig­ures presen­ted by the Grand Défi focused on arti­fi­cial intel­li­gence, only 10 to 15% of AI-based proofs of concept (in the sense of a demon­strat­or) in France are indus­tri­al­ised and go to scale. 

AI can pro­cess large volumes of com­plex data to speed up decision mak­ing at all levels of com­mand, act­ing as a force mul­ti­pli­er for com­mand – par­tic­u­larly with the emer­ging require­ments of multi-domain oper­a­tions. While human resources can cur­rently pro­cess, at best, 20% of the inform­a­tion pro­duced today, this per­cent­age may drop to as little as 2% in the face of the explo­sion in data pro­duc­tion³.

As a sign that this use case is of major interest to the armed forces, the Min­istry of the Armed Forces has just signed a defence and secur­ity con­tract worth €240 mil­lion, over sev­en years, with a non-tra­di­tion­al play­er in the defence industry: a pure play­er in defence AI in France. The aim of this con­tract is to accel­er­ate the intel­li­gence cycle and to pro­cess the tsunami of data com­ing from these sov­er­eign CSO optic­al ima­ging satellites. 

In addi­tion to hand­ling the expo­nen­tial growth of data, arti­fi­cial intel­li­gence can also improve the qual­ity of decision-mak­ing pro­cesses and become a key play­er in the plan­ning and con­duct of oper­a­tions con­tinuum. Indeed, the aggreg­a­tion of many data sources, and their ana­lys­is by machine learn­ing algorithms, can help determ­ine the best pos­sible geo­graph­ic dis­tri­bu­tion of forces based on the mis­sion, a unit’s cap­ab­il­it­ies, con­di­tions in the area of interest, resup­ply require­ments and inform­a­tion derived from the ana­lys­is of all intel­li­gence sources.

Every theatre of oper­a­tions offers an infin­ite num­ber of force con­fig­ur­a­tions. To get an idea of the mag­nitude of this num­ber, con­sider a game of chess. After each play­er has made four moves, there are 988 mil­lion dif­fer­ent pos­sible con­fig­ur­a­tions. Com­bin­ing massive sources of rel­ev­ant inform­a­tion (such as loc­a­tion data, weapon ranges, intel­li­gence, etc.) gives AI a crit­ic­al role in oper­a­tion­al plan­ning and decision-making.

AI can go bey­ond human cap­ab­il­it­ies and dra­mat­ic­ally reduce the response times of defence sys­tems to attacks by fast-act­ing weapons sys­tems (hyper­son­ic mis­siles, cyber-attacks or dir­ec­ted energy weapons) to bring about a rad­ic­al change in cap­ab­il­it­ies. Again, there are non-tra­di­tion­al play­ers in the defence industry, such as the US com­pany Andur­il, one of the few US start-ups to have suc­cess­fully scaled up. Its algorithms on board drones for tar­get­ing pur­poses enable the US army to accel­er­ate the decision loop at a tac­tic­al level and to get closer to real time to counter enemy man­oeuvres much more effectively.

On so-called broad-spec­trum effects, AI can also help to obtain a broad­er know­ledge and under­stand­ing of the bat­tle­field, and for example anti­cip­ate attempts to manip­u­late cit­izens. Indeed, AI-driv­en ana­lys­is of the effect on pub­lic sen­ti­ment of a kin­et­ic strike on infra­struc­ture, or AI-based solu­tions to detect dis­in­form­a­tion cam­paigns, are among the most advanced intel­li­gence ser­vice tools for digit­al influ­ence and inform­a­tion­al warfare. 

Con­crete applic­a­tions of AI for the armed forces do exist, and not all use cases of this “enabling tech­no­logy⁴” have yet emerged at this stage. Non-tra­di­tion­al play­ers in the defence sec­tor stand out as sup­pli­ers of these AI-based solu­tions, thanks to their abil­ity to deal with spe­cif­ic “user” needs and to pro­duce demon­strat­ors in “agile” mode based on exper­i­ments. How­ever, these play­ers are faced with the more gen­er­al dilemma of adopt­ing gen­er­al-pur­pose tech­no­lo­gies: to what extent and at what speed do these play­ers suc­ceed in chan­ging the tra­di­tion­al forms of organ­isa­tion of the armed forces to enable them to bene­fit from the advances provided by gen­er­al-pur­pose technologies?