2023-2043年科学技术趋势：跨越物理、生物和信息领域报告（卷2）（英文版）（286页）.pdf

《2023-2043年科学技术趋势：跨越物理、生物和信息领域报告（卷2）（英文版）（286页）.pdf》由会员分享，可在线阅读，更多相关《2023-2043年科学技术趋势：跨越物理、生物和信息领域报告（卷2）（英文版）（286页）.pdf（286页珍藏版）》请在三个皮匠报告上搜索。

1、Science&TechnologyTrends 2023-2043Across the Physical,Biological,and Information DomainsNATO Science&Technology OrganizationVOLUME 2:AnalysisScience and Technology OrganizationThe research and analysis underlying this report and its conclusions were conducted by the NATO S&TOrganization(STO)drawing

2、upon the support of the Alliances defence S&T community,NATO AlliedCommand Transformation(ACT),and the NATO Communications and Information Agency(NCIA).This report does not represent the official opinion or position of NATO or individual governments butprovides considered advice to NATO and Nations

3、leadership on significant S&T issues.DISCLAIMERDale F.Redinglvaro Martn BlancoAngelo De LuciaCol Laura A.Regan,USAF,PhDDaniel BaylissNATO Science&Technology OrganizationOffice of the Chief ScientistNATO HeadquartersB-1110 BrusselsBelgiumhttp:www.sto.nato.intDistributed free of charge for information

4、al purposes;hard copies may be obtained on request,subject to availability,from the NATO Office of the Chief Scientist.The sale and reproduction of this report for commercial purposes isprohibited.Extracts may be used for bona fide educational and informational purposes,subject to attribution to the

5、NATO S&T Organization.Unless otherwise credited,all non-quantitative graphics were generated through deep learning,text-to-image Gener-ative AI models using Stable Diffusion 2.1 or sourced from and used under Creative Commonslicensing.Copyright NATO Science&Technology Organization,2023First publishe

6、d,March 2023ForewordAstheworldslargestcollabo-rative forum forscience and tech-nology(S&T)re-search in securityand defence,theSTO aims to em-powerNATOstechnological edge.It works to achievethis vision through an active network of around 5,000dedicated scientists,engineers,and analysts who carryout m

7、ore than 300 research activities annually.Research-ing S&T trends,developing new forecasting method-ologies,and identifying novel and disruptive S&T areall critical aspects of this work.As a community ofS&T experts,the STO collaborative network providesunparalleled insights into the latest trends in

8、 defence andsecurity technology,as well as their future evolution andimpact.This second volume of Science&TechnologyTrends 2023-2043 highlights this networks methods,data,and insights on emerging and disruptive technolo-gies,providing a deep dive into the foundations uponwhich Volume 1 of this repor

9、t is built.Mr.John-Mikal StrdalDirector,NATO STO Collaboration Support OfficeForecasting the de-velopment of S&Tpotentiallyrele-vant to NATO overa twenty-year pe-riod is daunting.However,throughrigorous and wide-ranging qualitativeandquantitativemethods,insightsinto the state,rate,and impact of S&Tr

10、elevantforde-fence and security can be constructed.Combininginsights from futures studies,STO technology watch,serious gaming,research meta-analysis,surveying expertopinion,and reviewing national research programmesprovides useful insights into S&T developments.Thisvolume summarises and synthesises

11、these activities,pro-viding a snapshot of future EDT development.It reflectsthe efforts of many individuals and 70 years of historyand intellectual richness of the STO as NATOs originalinnovation engine.Dr.Catherine WarnerDirector,NATO STO Centre for Maritime Research andExperimentationForeword cont

12、.Properly assessingthe state,rate,andimpact on the Al-liance of EDTs andotherpromisingtechnologiesre-quires an evidence-based understand-ing of the over-arching scientificand technologicalecosystem,includ-ing weak technology signals.Various data sources andtechniques were employed to support the ana

13、lysis under-taken in this report.Over four thousand articles,books,meta-studies,documents,and reports were collected,collated,reviewed,and assessed.Supplementing thiscollection,the NATO Office of the Chief Scientist andthe NATO Communications and Information Agency(NCIA)developed the Science&Technol

14、ogy EcosystemAnalysis Model(STEAM).STEAM draws upon a grow-ing collection of millions of English-language journalarticles,pre-prints,and abstracts.STEAM allows deeperinsights into EDTs and creatively exploits ArtificialIntelligence to better understand developments withinthe S&T ecosystem,ultimately

15、 putting the analysis on astronger footing.Nevertheless,STEAM is still a work inprogress,with its analysis more indicative than definitive.With an eye towards the next S&T trends assessment,plans are being made to expand STEAM to includepatent data,non-English language articles,and deeperAI-enabled

16、analysis of S&T trends.In addition,theplanned role of STEAM will be enlarged into a broaderbusiness intelligence tool.This will enable a deeperunderstanding of international research and capabilitydevelopment collaboration patterns while supportingNATOs S&T portfolio management.In turn,this willallo

17、w NATO to anticipate,deliver and react to EDTdevelopments and enabled capabilities.Mr.Dale F.RedingSenior Scientific AdvisorNATO STO Office of the Chief ScientistTable of ContentsForeword.iiiExecutive Summary-Volume 2.vii1Introduction.11.1Context11.2Analysis21.3Approach21.4Overview32Methodology.42.1

18、Description42.2Assessment42.3Information Sources102.4Alliance and Partner Research Programmes182.5MetaStudies,Articles and Reports213EDT Analysis.223.1Description223.2Exploration of Selected EDTs234Conclusion.24Appendices.25AArtificial Intelligence.26BRobotics and Autonomous Systems.44viCBiotechnolo

19、gy&Human Enhancement.57DData.75EElectronics and Electromagnetics.100FEnergy&Propulsion.113GHypersonics.125HNovel Materials and Manufacturing.135IQuantum Technologies.149JSpace Technologies.164KSTO Technical Network Survey.179LS&T Ecosystem Analysis Model(STEAM).201Bibliography.209Symbols,Abbreviatio

20、ns and Acronyms.272Executive Summary-Volume 2Science&Technology Trends:2023-2043 provides an updated assessment of Science&Technology(S&T)trends and their potential impact on NATO military operations,defence capabilities,enterprisefunctions,and political decision space.Such an assessment draws upon

21、the collective insights andresearch activities of the NATO Science&Technology Organization(STO),its collaborative network ofover five thousand active scientists,analysts,researchers,engineers,and associated research facilities.These insights have been combined with an extensive review of the open-so

22、urce S&T literature,selectednational research programs,NATO STO technology watch activities,(serious)research games,STOCPoW(Collaborative Programme of Work)activities,and NATO innovation endeavours.Given the volume of work undertaken,the report has been split into two volumes.The first sum-marises t

23、he analysis and discusses,in general terms,key geopolitical and social trends impacted by S&Tdevelopments or the impact of those developments themselves.The second provides a detailed discussionof the methodology,data,and analysis that underpins the recommendations and observations in Volume 1.This

24、second volume also provides a deep dive into the individual emerging and disruptive technologies,their potential impact on military capabilities,and their impact on other EDTs.Extensive use is madeof quantitative analyses of recent academic articles,scientometric assessments,and a survey of the STOn

25、etwork to assess the state,rate,and potential impact of these technologies.Extensive references are alsoprovided to support the conclusions and facilitate an even deeper analysis by the reader.1.Introduction“We are still the masters of our fate.Rational thinking,even assisted by any conceivable elec

26、troniccomputors,cannot predict the future.All it can do is to map out the probability space as it appears atthe present and which will be different tomorrow when one of the infinity of possible states will havematerialised.Technological and social inventions are broadening this probability space all

27、 the time;it isnow incomparably larger than it was before the industrial revolutionfor good or for evil.”DennisGabor 1The Future-Mapping Out the Probability Space1.1ContextSince the last Science and Technology(S&T)trends assessment 2,the world has seen technological,social and geopolitical change an

28、d disruption at an unprecedented level.Over the last three years,NATO,and the world as a whole,have been challenged by the pandemic of the century(COVID-19),theRussian-Ukrainian war,the withdrawal of NATO forces from Afghanistan,significant climate disruption,the rise of potential nuclear powers,inf

29、lation,increased tensions in the Asia-Pacific,the challenge oftechno-authoritarianism,as well as insatiable S&T progress,especially in the areas of Data,Quantum,Biotechnology,Autonomy and AI.NATO is a unique consultative and collaborative military and political framework.This collaborationextends be

30、yond the military,and political realms to science and technology,enabling the application ofstate-of-the-art validated knowledge for defence and security purposes.NATO S&T activities embracescientific research,technology development,analytical assessments,capability planning,experimentation,and a wi

31、de range of related scientific activities 3.The Science and Technology Organization(STO)plays a decisive role in supporting innovation;providing profound insights into alliance challenges;ensuring the integration of Alliance capabilities;and making available an interconnected network of science and

32、knowledge workers capable of providingevidence-based advice to NATO,as well as alliance members and partners.At its core,the role of NATOsS&T community is to 3:“.maintain NATOs scientific and technological advantage by generating,sharing and util-ising advanced scientific knowledge,technological dev

33、elopments and innovation to supportthe alliances core tasks.”To support the strategic development of such technologies,the STO is responsible for assessing S&Tfutures for senior leadership.As noted in the STO charter(2012)3:1.2 Analysis2“To fulfil its mission,the STO will.provide advice to NATO and

34、Nations leadershipon significant S&T issues,including the identification of emerging technologies,and theassessment of their impact on defence and security.”1.2AnalysisScience&Technology Trends(2023-2043)provides context and a foundation for a NATO technologystrategy,Alliance capability development

35、and the NATO S&T programmes of work.The core objective isto increase the understanding within the Alliance of the potential for S&T developments to enhance orthreaten Alliance military operations.Increased awareness of the state,the projected rate of development,and the anticipated impact of such te

36、chnologies should lead to enhanced future capabilities,better strategicdecision-making and improved S&T portfolio management.Such an assessment does not attempt topredict the future in detail(a difficult task at best and impossible at worst).Instead,it seeks to provide acontext for anticipating the

37、possible development of S&T and its potential impact on the Alliance.Analyses of technology trends and the associated process of technology watch are critical steps toidentify new militarily important technologies and communicate the potential impact of these technologieson NATO and national leaders

38、hip.Those recognised technologies hold the promise to enable thedevelopment of disruptive military capabilities for Alliance(BLUE)and potential adversarial(RED)forces.The report assesses S&T trends(emerging and disruptive technologies)projected to impact NATOoperations,capability development and cor

39、e functions over the next 20 years and explores the implicationsof these changes.These S&T areas are broad,have significant overlaps,are not orthogonal and areexpected to:Mature over 20 years;Be transformative or revolutionary;and,Be emergent or create generational shifts in S&T development.1.3Appro

40、achThis report aims to reach a wide audience both inside and outside of NATO and its partners.We do soto stimulate a frank and open discussion of potential opportunities and risks presented by technologicaldevelopments over the next 20 years.As such,the report is based strictly on the following:Tech

41、nology trends discussed in the open literature;A global perspective on technological progress;Scientometric analysis;and,Logical reasoning informed by Alliance S&T expertise and technology watch activities.Defence Ministers approved a canonical set of EDTs and an associated roadmap in October 2019.I

42、n2022,the 2022 Alliance Heads of State and Government Madrid summit authorised two more prioritytechnologies for consideration,Novel Materials and Agile Manufacturing and Energy and Propulsion.Inkeeping with the STOs mandate to continue to monitor and evaluate the broader technological landscape,thi

43、s report also considers the status of recent developments in Electronics&Electromagnetic(E&EM)technologies,including developments in directed energy weapons.As such,10 EDTs are considered indetail in this report,each broken down into technology focus areas,highlighting specific areas of researchand

44、development(R&D).Chapter 2 discusses this decomposition in further detail.Science&Technology Trends:2023-2043 supersedes the Science&Technology Trends:2020-2040report 2 but draws heavily upon its foundations,structure,insights and lessons learned.Indeed,thisreport is best considered an update to the

45、 previous report.As before,the report exploits a broad range3Chapter 1.Introductionof open-source reports,internal assessments,NATO EDT and innovation activities,technology reviews,serious games,quantitative analysis and futures studies to develop a comprehensive understanding of thefuture technolog

46、y landscape.These sources include:Existing NATO S&T trend and future security environment studies,strategies,discussions andassessments;Technology watch activities conducted by the STO,including existing Technology Watch Cards(TWC)(current as of October 2022),Chief Scientist Reports,(serious)technol

47、ogy games and VonKrmn Horizon Scans(vKHS);Meta-analyses and reviews of open source technology watch and futures research from defence,security and industry sources;Internal and external quantitative analysis of academic publications,patents and research activities;Scientometric analysis of the globa

48、l S&T ecosystem;Surveys of the STO network and panels seeking insights into technological developments,readinessand maturity;NATO-sponsored EDT workshops and innovation system engagements;and,Alliance and partner EDT studies and research programs.1.4OverviewWithin the following appendices,an analysi

49、s is presented of identified and militarily relevant S&T trends,which may impact NATO capability development and operational challenges over the upcoming 20 years(2023-2043).The approach and key data sources used to conduct this assessment are described in Chapter2.Separate appendices provide a more

50、 detailed exploration of each EDT,drawing heavily upon STOresearch and technology watch activities.This section also includes Conjecture Cards,short vignettes thatdescribe the potential future application of these technologies.They are included to help contextualisethe potential impact of these tech

51、nologies but do not necessarily represent research interests or activitiesbeing undertaken by NATO or members of the Alliance.The bibliography at the end of this document provides an extensive list of useful references.Theseare also used throughout the body of the text where appropriate.When using t

52、he Adobe PDF version ofthe report,clicking on a numbered reference will take the reader to the relevant entry in the bibliography.If desired and available,clicking on the provided URL(i.e.,web-link)will allow the reader to open thesource reference directly for further study and exploration of the to

53、pic.2.Methodology“Finding patterns is easy in any kind of data-rich environment;thats what mediocre gamblers do.Thekey is in determining whether the patterns represent noise or signal.”-Nate Silver 4Forecasting:2.1DescriptionForecasting S&T development is a challenging task that demands a structured

54、,thoughtful and information-intensive process.We have taken such an approach in developing this report,drawing from multiplesources and assessments while seeking common patterns or indicators from within the informationavailable.For transparency and challenge,the approach and key data sources used t

55、o conduct thisassessment are described in the following sections.2.2AssessmentTo understand the state,rate and impact of EDT development,it is necessary to consider the following:technological maturity(current and future);keywords defining sub-areas and other technological connections;level of atten

56、tion or hype around a particular technology or scientific area;integration into NATO operational and enterprise capabilities;and,potential military impact.Such an assessment is problematic as each EDT encompasses many different core aspects,eachpotentially at a different stage of development.As a re

57、sult,for this report,each EDT is broken into asecond level of areas identified for focused development or research.We refer to these as EDT technologyfocus areas.These areas are very broad,and assessments of the state and rate of development willnecessarily be somewhat fuzzy in nature.The report emp

58、loys several approaches described in more detailbelow.5Chapter 2.Methodology2.2.1Technological MaturityIn general successful S&T proceeds along a developmental path captured as technology readiness levels(TRLs),originally developed by NASA 5,6.Each level is a potential off-ramp or pause for that par

59、ticulartechnology(see Table 2.1).These levels provide a useful shorthand for interpreting technology maturityand,as such,are widely used within industry and government.Similar frameworks 7,8,9 are availablefor human,algorithm,manufacturing,commercialisation,machine learning,and technology commitment

60、readiness levels.Table 2.1:Technology Readiness Levels.TRL 9Actual system proven through successful mission operations.TRL 8Actual system completed and qualified through test and demonstration.TRL 7System prototype demonstration in a space environment.TRL 6System/subsystem model or prototype demonst

61、ration in a relevant environment.TRL 5Component and/or breadboard validation in a relevant environment.TRL 4Component and/or breadboard validation in a laboratory environment.TRL 3Analytical and experimental critical function and/or characteristic proof-of-concept.TRL 2Technology concept and/or appl

62、ication formulated.TRL 1Basic principles observed and reported.Technologies,or the underlying sciences,may be unsuccessful in generating new operational capabili-ties or pause at any point along the path to TRL 9+due to several factors,including lagging dependencieson other technologies,costs,ethics

63、,policies or fundamental physical,information or human limits.Eval-uating when these technologies will reach TRL 9 assesses full technological maturity and utility foroperational and enterprise capabilities.Unsuccessful developments are necessary within S&T to inform and ultimately give rise to newa

64、pproaches and technologies that may be more successful in moving from basic principles to operationalcapabilities.As the American inventor Thomas Edison stated:“I have gotten a lot of results!I knowseveral thousand things that wont work.”10.Thus,constructive failure is a feature of scientific andtec

65、hnological developments,not a bug.We assess the current TRL levels primarily through STO Technology watch activities assessments,asurvey of the STO experts network,related futures assessments,and references to several available TRLcalculators 11.It should be noted that emerging technologies are usua

66、lly in the range of TRL 1 through5 12.2.2.2KeywordsKeywords associated with an EDT provide insights into research areas,sub-areas and synergies.Keywordsfor this report are drawn from several quantitative and qualitative sources.They include analysis ofpublication databases,publication guidelines(e.g

67、.IEEE 13),critical technology lists 14,15,16 andSTO survey results.2.2.3AttentionTechnological development is distinctly cyclic on many levels.The most well-known of these cycles isthe Gartner Hype Cycle 17(Figure 2.1),itself based on Howard Fosdicks work on the sociology oftechnology adoption 18,19

68、.While this relates to TRLs,such an assessment reflects a socio-technicalperspective on the state of development of technology and the likelihood of continued advancements inthis area.Essentially this captures buzz,or more accurately attention around a technology.Technologies do not always progress

69、from the beginning to the end of such a cycle;indeed,as notedearlier,most technologies fail.Moreover,many avenues of science or technological discovery never breakthrough to ignite innovation.Instead,they disappear from public consciousness after initial enthusiasmas unproductive avenues of developm

70、ent,or they may appear later on as new convergent developmentsreinvigorating an old idea.Finally,even successful technologies may reappear as novel ideas create2.2 Assessment6innovation triggers and old technologies become so integrated into production systems that the originalconnection is lost on

71、all but the most technically minded.Such an evolutionary process built on heroicfailures 20 or creative errors is essential to scientific and technological progress,as lessons and ideasthat arise will often lead to entirely new areas for exploration,innovation and development.During a hype cycle,a s

72、uccessful trending technology will(arguably)ultimately go through five keyphases:21,22:Innovation Trigger:After considerable supporting research,a potential new technology break-through shows promise.This initial innovation trigger builds upon early experimentation and resultsin proof-of-concept sto

73、ries,and media interest is triggered.This spark yields growing publicity andinternet search activity.However,no viable product exists at this stage,and commercial viabilityremains unproven.Peak of Inflated Expectations:Early publicity produces many success stories often accompa-nied by scores of fai

74、lures.Interest(e.g.as measured by web searches)is at an all-time peak.Someinnovative companies take action;many do not.Trough of Disillusionment:The limitations of the technology become clear,and some implemen-tation efforts fail to produce useful results.As a result,general interest falls,and negat

75、ive storiesbecome more frequent,although these may be overly pessimistic.Eventually,some developersand producers move on to other areas or fail outright.A bifurcation occurs at this point,whereinvestment and continued developments only occur if continued progress can be shown throughthe refinement o

76、f the underlying technology,the development of a better understanding of wherethis technology is most applicable or a convergence of other technologies or demand.If this doesnot happen,the technology will eventually be deemed unproductive and disappear entirely fromconsideration,or return to the sta

77、rt gate to await further developments,technological convergenceor changing circumstances.Slope of Enlightenment:With a better understanding of what is practical and where it can be bestapplied,the potential begins to crystallise and become more widely understood and appreciated.Next-generation produ

78、cts occur,and positive attention increases with more successful trials andpilot products.Some companies remain cautious.Plateau of Productivity:Mainstream adoption occurs.With a better understanding of value,applicability and limitations,the technology has found its market.Issues and new ideas may s

79、tillarise,potentially kicking off a new cycle.Otherwise,the technology becomes so well integrated intothe technological landscape that its use becomes commonplace until a new technological advancesupplants it.Balancing out the highs of inflated expectations and the lows of the trough of disillusionm

80、ent is criticalin making investment and long-term capability decisions.The exact placement of technology on such acurve is problematic,as is the focus on hype rather than a measurable quantity,such as attention.Thisreport will distinguish between emerging technologies(those in the early steps of the

81、 curve)and disruptivetechnologies(typically towards the middle to end of the Gartner curve).In contrast to the last S&T report 2 this report does not track hype trends,although the Gartnerannual report 23 was consulted,along with many other sources.This decision was based on a beliefthat the hype as

82、sessment added little to the discussion for NATO above and beyond the TRL assessment.Figure 2.1 presents the Gartner consolidated assessment for all emerging technologies considered tobe of interest to a general business audience as assessed in 2022.While this is very useful as a sourceof data and a

83、nalysis for understanding EDTs,it does not focus on defence applications and considersdevelopments over a 2 to 10-year horizon.One of the more interesting observations in the 2022 emerging technologies list is the identificationof three major themes:evolving/expanding immersive experiences,accelerat

84、ed AI automation,and7Chapter 2.MethodologyFigure 2.1:Gartner Hype Cycle 2022(CREDIT:Gartner 24).2.2 Assessment8optimised technologist delivery.When explored in detail,these themes align well with this reportsidentified technology themes(intelligent,interconnected,decentralised,and digital),which wer

85、e carriedover from 2.Figure 2.2 presents the Gartner assessment for sub-areas associated with AI,considered to be ofinterest to a general business audience as assessed in 2022.Figure 2.2:Gartner Hype Cycle-Artificial Intelligence 2021(CREDIT:Gartner 25).2.2.4CapabilityFor NATO,EDTs are primarily of

86、interest through their influence on current and future Alliance defencecapabilities.To better connect EDTs to their military impact,each EDT is evaluated for its potential effecton NATO operational capabilities.NATOs operational capability taxonomy 26 provides a structured listof capabilities and su

87、b-capabilities.In addition,the first level capability Enable is added to this list tocapture NATO enterprise functions unrelated to a specific military operation.The first level capabilitiesused are:ENABLE:This capability supports the strategic or organisational functions necessary to supportAllianc

88、e political and military operations in all domains,levels,and across all MCAs.For ourpurposes,Enable includes additional basic business processes such as Govern,Generate,andOperate,identified in 26.These are the capabilities required to run and manage any large defenceand security establishment.Such

89、 generic processes focus on support services required throughout9Chapter 2.Methodologythe defence enterprise,such as financial,human resources,policy development,defence planning,legal,etc.,interfacing with almost every other process in the enterprise,whether governance,strategy,force development,or

90、 operations.PREPARE:To establish,prepare and sustain sufficient and effective presence at the right time,including building up forces through appropriate and graduated readiness,to meet any requirements,keeping enough flexibility to adapt to possible changes in the strategic environment.These alsoin

91、clude the capabilities to contribute to Deterrence and Defence,Resilience and Projecting Stability.PROJECT:To conduct strategic deployment of headquarters(both for the NATO Force Structureand at a national level),forces and capabilities supporting any Alliance mission.These also includethe capabilit

92、y to contribute to deterrence.SUSTAIN:To plan and execute the timely support and sustainment of forces,including essentialmilitary infrastructure,movement and transportation,military engineering support,contracting,supply,maintenance,services management,basing support and health and medical support.

93、ENGAGE:To perform the tasks that contribute directly to achieving mission goals within collectivedefence,crisis management,and cooperative security operations.It includes all capabilities requiredto defeat,if necessary,adversaries as well as other capabilities such as,among other things,thoseneeded

94、to evacuate non-combatants,prevent the use of force by opponents,train local securityforces and participate in stabilisation and reconstruction.PROTECT:To conduct strategic deployment of headquarters(both for the NATO Force Structureand at a national level),forces and capabilities supporting any All

95、iance mission.These also includethe capability to contribute to deterrence.INFORM:To establish and maintain the situational awareness and knowledge required to allowcommanders at all levels to make timely and informed decisions.CONSULT,COMMAND,AND CONTROL(C3):For commanders to exercise authority ove

96、rand direct the full spectrum of assigned and attached forces in accomplishing the mission.Includethe capability:to communicate and coordinate with other actors which are present or involved inthe operational area and effective information exchange with the political and military leadership;the capa

97、bility to plan,employ and coordinate civilian activities with other actors and organisations;capability for nuclear planning and political consultation that allow the rapid development of nuclearemployment options in crisis and war,should circumstances so dictate.An assessment is presented for Enabl

98、e and the first level of the operational taxonomy only:Prepare,Project,Engage,C3,Sustain,Protect,and Inform.This assessment is presented later in Appendices A-J.2.2.5ImpactAssessing the potential impact of emerging or disruptive technologies is a complex process.To doso successfully requires conside

99、ration of the threat environment(current and future),legal&policyconstraints,political factors,and investment decisions,as well as estimating the potential for organisationaluptake(e.g.entrepreneurial drive and risk tolerance)27.These estimates are further compounded if theroad to disruption involve

100、s complex combinations of such technologies(e.g.synergies)or requires thedevelopment of new concepts.For purposes of this report we follow 27,defining Impact in a somewhat subjective and imprecisemanner as(Table 2.2):2.3 Information Sources10Table 2.2:EDT Impact.ScalePerformance:speed,range,accuracy

101、,lethality,survivability,affordability,availability,de-pendability or other defining capability characteristicModerate10-50 improvement%High50-100 improvement%RevolutionaryGreater than 100%,or conducting activities or tasks hitherto deemed impractical or impossibleAssessments in this report of the m

102、ilitary impact of relevant technologies are based predominately onthe results captured in 2 and a wisdom-of-the-crowd”assessment,supported by a survey of the NATOSTO network.An in-depth discussion of the survey results may be found in Appendix K.2.3Information Sources2.3.1NATO Reports and StudiesThe

103、 following NATO-released documents were used in the preparation of this report:Framework for Future Alliance Operations(2018)In future operations,NATO must continually evolve,adapt,and innovate to maintain a decision andcapability advantage(credible,networked,aware,agile and resilient).The Framework

104、 for Future AllianceOperations 2018(FFAO)28 provides a futures perspective supporting such developments.It informs theAlliance of opportunities to improve its defence and deterrence posture and its ability to project stability,ensuring it remains continuously proactive,ready and responsive.It descri

105、bes how NATO forces can keepthe edge and retain the ability to defeat potential adversaries on future battlefields.Finally,it providesmilitary advice identifying force characteristics and abilities the Alliance needs to retain the military edge,address upcoming challenges,and seize future opportunit

106、ies.The FFAO identifies several instability situations,potential events of critical significance,which couldreach the threshold requiring the Alliance to use military forces.These instability situations provide auseful framework for assessing the impact of EDTs,both from a threat and opportunities p

107、erspective,andare listed below:Weapons of Mass DestructionConventional WarThreat EscalationHybrid WarIrregular WarTerrorismGlobal Commons DisruptionCritical Infrastructure AttackInformation WarfareCyberattackGovernance ChallengesEndangerment of Civilian PopulationsMass MigrationPandemic DiseaseNatur

108、al or Man-made DisasterThe report identifies anticipated future operational challenges.These include the impact of technologi-cal advances;new concepts of operation(e.g.global strike,hybrid,and cyberspace operations);and shiftsin the geopolitical landscape.Of interest to the assessment of EDTs,the r

109、eport notes that future-armedconflict is expected to be characterised by any combination of:Adversaries(state and non-state)global inscope and employing indirect approaches;A greater role of super-empowered individu-als and non-state actors that produce hard-to-predict effects;A compression of strat

110、egic,operational andtactical decision-making,blurring decision-making processes;More inter-connectivity across air,land,sea,cyber,space and information domains;Small units fighting over greater distances;Operations in the cyberspace domain,globalcommons,urban areas,and subterranean ar-eas;Rapidly em

111、erging and widely available tech-nologies;11Chapter 2.MethodologyThe use of human enhancement and the risingimportance of the human-machine interface;The use of automated and potentially au-tonomous systems and operations in whichhumans are not directly involved in the deci-sion cycle;New classes of

112、 weapons that can causewidespread destruction;Greater number of sensors and the prolifera-tion of the internet of things;An expanded access to knowledge,includingthe ability to conduct large-scale advanceddata analytics to gain a military advantage;and,Weaponised information activities intendedto in

113、fluence populations alone or in supportof armed conflict.Emerging or Disruptive Technologies Roadmap(2019)NATO Defence Ministers approved an EDT Roadmap in October 2019.This provided a canonical list ofseven EDTs providing structure to subsequent innovation,strategies and associated roadmap developm

114、entswithin NATO over the last three years(e.g.29,30.Following this,in February 2021,NATO DefenceMinisters approved an EDT strategy to guide NATOs development of EDT policy 31.Science&Technology Trends:2020-2040(2020)This report 2,published in March of 2020,highlighted S&T trends assessed by the NATO

115、 STO.It was thesecond comprehensive report on emerging trends in S&T by the NATO STO,with the first being publishedin 2017 32.The report exploited a broad range of open-source reports,internal assessments,and studiesof potential futures to develop a comprehensive understanding of the future technolo

116、gy landscape.Thesesources included:1)Existing NATO S&T trend and future security environment studies,discussionsand assessments;2)Technology watch activities conducted by the S&T Organisation,including existingTechnology Watch Cards(TWC)(current as of Feb 2019)and von Krmn Horizon Scans(vKHS);3)Meta

117、-analyses and reviews of open-source technology watch and futures research articles/reports fromdefence,security,and industry sources;4)NATO-sponsored EDT workshops and innovation systemengagements;and,5)Alliance and partner EDT studies and research programs.The EDT taxonomy used in the 2020 S&T Tre

118、nds report 2 builds on the canonical set of seven EDTs,adding Materials(Novel Materials and Agile Manufacturing)based on STO foresight activities.DefenceMinisters approved the first seven EDTs in October 2019,while the STO added an eighth(Materials)area for future consideration and development.Inclu

119、ding an eighth EDT recognises that materials andmanufacturing research are well-developed technologically yet increasingly disruptive(e.g.3D/4Dmanufacturing).At the same time,some aspects are emergent(e.g.novel materials,bio-manufacturing,and nanotechnologies).These eight highly interrelated S&T are

120、as are still forecasted to be major strategic disruptors over thenext 20 years.These S&T areas are either currently in nascent development stages or undergoing rapidrevolutionary development.The EDTs identified were:DataArtificial Intelligence(AI)AutonomySpaceHypersonicsQuantumBiotechnologyMaterials

121、2.3.2NATO STO Technology Watch ActivitiesCollaborative Research Program(CPoW)NATO S&T Priority AreasA set of NATO research priorities guides S&T conducted under the auspices of the STO,as agreed to bythe Nations through the S&T Board(STB).These priorities influence medium to long-term S&T planningac

122、ross NATO and inform S&T investment decisions within the Nations.In addition,the STO maintainsan understanding of current and future S&T,including broad themes and EDTs,through engagement withthe approximately 5000 active scientists,engineers and analysts participating in the collaborative researchp

123、rogram(CPoW).The STO S&T priorities are organised into ten S&T Areas,spanning the Human,information andphysical sciences.Each area has a specific defined Targets of Emphasis(TOE).While there are many2.3 Information Sources12ways of organising S&T activities,the ten S&T Areas provide a broad and usef

124、ul reference frame forresearch activities,while the TOEs provide selective focus and orientation.The priorities are constructedindependent of physical domains,scientific disciplines,or specific applications,while the language issituated between the words used to express requirements and S&T solution

125、s.The employment of theprioritiesfocuses onS&Tefforts thatsupportinnovativecapabilitiesfor theAlliancesforces,informfuturemilitary specifications,and provide strategic advice to senior decision-makers.Table 2.3 summarisesthese priorities and associated targets of emphasis.Table 2.3:NATO S&T Prioriti

126、es.DomainNATO S&T Priority AreasTargets of EmphasisHUMANAdvanced Human Performance&HealthMedical Solutions for Health OptimisationHuman ResiliencyEnhanced Cognitive PerformanceHuman&Machine InterfacesCultural,Social&Organisational BehavioursSocial InfluencePolitical InfluenceCultural CommunicationsG

127、roup&Organisational BehaviourINFORMATIONData Collection and ProcessingEM SensorsNon-EM SensorsSensor Integration&NetworksAdvanced Signal ProcessingInformation Analysis&Decision SupportBig Data&Long Data Processing and AnalysisBig Data&Human Decision MakingMulti-Domain Situational AwarenessPlanning a

128、nd Managing UncertaintiesAdvanced Systems ConceptsIntegrated Human-Machine Hybrid ForceClusters&SwarmsModular,Scalable SystemsHigh Assurance Engineering&ValidationAutonomyArtificial IntelligenceMission Autonomous SystemsHuman-Autonomous Machine TeamingCommunications&NetworksSecure and Resilient Comm

129、unicationsTrusted Multi-Domain Information SharingAd hoc and Heterogeneous NetworksPHYSICALPrecision EngagementPrecision ControlWeapons-Techniques and SystemsWeapons-EffectsActive&Passive EM,Acoustic&Optical CountermeasuresPlatforms&MaterialsFast and Agile PlatformsUnmanned PlatformsHypersonic Platf

130、ormsAdvanced and Adaptive MaterialsIn-Theatre Fabrication&Production of EquipmentPower&EnergyPower&Energy StorageAlternative&Renewable Energy SourcesPropulsionEnhanced Energy Efficiency&ManagementTargets of emphasis do not naturally align with the more broadly identified EDTs nor provide sufficientr

131、esolution of potential or current development areas found within the identified EDTs.Nevertheless,theyare useful for providing insights into developing a second layer to the EDT decomposition.For purposesof this report,we refer to these as Technical Focus Areas(TFA).TFAs are sub-aspects of EDTs suit

132、ablefor focused research.Technology Watch Cards(TWC)Recognising the pressing need to maintain the Alliances technological edge,the STO actively pursuesTechnology Watch for the Alliance.The STO Panels and Group have embraced a culture of continuallyidentifying and documenting potentially disruptive s

133、cience or technology in Technology Watch Cards.These cards contain assessments of the maturity of the science or technology and offer commentary onhow science or technology may affect the capabilities of the Alliance and potential adversaries in the13Chapter 2.Methodologyfuture.The current S&T Trend

134、s report relies heavily on the almost one hundred Technology Watch Cardsdeveloped by the STO Panels and Group to deliver a short synthesis of observed technology trends.TWCassessments and text were especially helpful in drafting the more detailed appendices in this report.TWCcards that have been rec

135、ently developed,updated or are in the process of being updated are:AVT:Hypersonic Vehicles Additive ManufacturingHFM:Digital twin of the human Emotional design;Designing interface that takes into account emotionalresponses Trust engineeringIST:Probabilistic Programming Languages Electromagnetic Meta

136、surfaces Free Space Optical Commu-nications-Communications and Networks Self-Organising Networks-Communications and Networks Compressive Sensing for EO-IR systems Neuroelectronics Intelligent Autonomy Blockchain Tech-nology Digital Twins Multi-Party Computation Real Avatars Quantum-safe cryptography

137、 QuantumTechnologies-Communications and Networks,Computing&Simulation,Sensing&Imaging Blockchainand Distributed Ledger Technologies SAS:(N/A)SCI:Alternative Computing Architectures Integrated Defensive Aid Systems AI&counter-AI SensingSystems System V&VSET:Cognitive EW Early warning passive RF syste

138、ms Fully digital RF Systems Quantum Sensing Neuromorphic Processing Crystal Vacancy Centre Sensors Metal-Organic Frameworks Space-borneHyper-spectral Capability Adaptive Optics Multi-band Thermal Imagery SWIR imagery Remote DigitalHolography Ultrashort Laser Application for Defence and Security Inte

139、grated RF Photonics DeepLearning for Military EO/IR ATRMSG:Modelling and Simulation for social media Artificial Intelligence-Data-driven Behaviour Modelling Immersive Simulation Devices for Improving Dismantled Soldier Reparation with Augmented and MixedReality DevicesAs most tech watch cards are NA

140、TO UNCLASSIFIED,they are used predominately to flag areasof interest and provide contextual information.All conclusions in this report are based on open-sourcematerials.Von Krmn Horizon Scans(vKHS)To address emerging challenges,the von Krmn Horizon Scans(vKHS)is an instrument to quicklyperform a tec

141、hnology scan on a particular S&T topic within an abbreviated period(typically two to 6months).The process assesses the state of leading-edge research in a specific S&T area,the outlook forthe next decade,its relevance for the armed forces,and potential avenues for investment.They draw uponinternatio

142、nally recognised S&T expertise and experienced senior military.von Krmn Horizon Scanshave been undertaken on laser weapons,quantum technologies,artificial intelligence,and optronic 3Dimaging systems.WorkshopsFrom 8 to 9 February 2021,the NATO Science&Technology Organization(STO)conducted theDisrupti

143、ve Technologies Table-Top Exercise(D3TX),its first fully virtual serious(public release)game toassess the military relevance and potential impact of emerging and disruptive technologies(EDTs)33.This ambitious event attracted over two hundred active participants from various backgrounds,includingthe

144、armed forces,policy-making,procurement,defence research,academia,and industry,from Allied andmany Partner Nations.Participants of different professional backgrounds were grouped into small syndicates of eight tofourteen players.Each syndicate played within one scenario but addressed all capabilities

145、 and assessed alltechnologies.Further,the D3TX provided a means for exploring the operational impact of previouslyidentified EDTs and TFA.However,it also identified innovative technology areas(so-called weak signals).These areas have been considered in the development of this report.The D3TX tackled

146、 the challenging assessment of the technological impact on military operationsfrom several different angles:2.3 Information Sources14Figure 2.3:Disruptive Technologies Table-Top Exercise(D3TX)33.The D3TX was played against four bespoke scenarios that covered the full spectrum of Alliancecore tasks a

147、s identified in NATOs Strategic Concept:Collective Defence,Crisis Management,andCooperative Security.The D3TX addressed the comprehensive set of required military and enterprise capabilities acrossthe air,sea,land,cyber,and space domain.The capabilities were:Enable,Prepare,Project,Sustain,Engage,Pro

148、tect,C3 and Inform.The D3TX assessed a broad range of emerging and disruptive technologies,represented by twenty-seven Technology Cards(based on 2),while also allowing players to identify“weak(technology)signal”cards.2.3.3SurveysFeedback on Emerging S&TDuring the spring STO panel meetings,the STO pa

149、nels were asked to provide a rough assessment ofwhich technology areas were considered emergent and disruptive.This allowed additional detail to beadded to the development of the TFAs,augmenting the information derived from the TWCs.The listgenerated per panel:AVT:Electric drive Renewable energy Sin

150、gle staged orbit(launch)Energy management HFM:Ethical,Legal and Social Implications(AI and BHET)IST:AI Big data analytics and visualisation Robotic process automation Quantum computing andalgorithms Virtual/Augmented/Extended(VR/AR/XR)reality,tele-existence and real avatars Data-centricarchitecture/

151、security Distributed resilient infrastructure:Tactical cloud/edge computing/HeterogeneousMANET/SDN/NFE for tactical networks/cognitive radio/cognitive networks Military Internet ofMilitary Things Counter/adversarial-AI and AI for cyber Homomorphic encryption,quantum-safe crypto Trust architectures/d

152、igital trust Ethical,legal,societal and environmental aspects of technology bydesign/human-machine teaming ICT(Information and Communication Technology)Supply chain security(devices and materials)15Chapter 2.MethodologySAS:Non-military/non-government leadership in tech Blockchain Swarm tech(cooperat

153、ive autonomy)Critical materials shortages/synthetic substitute.State of fear Data science Complexity and uncertainty Invisibility(materials)Neuroscience,mind control Hypersonic/autonomy combinedSCI:Alternative Computational Architectures Quantum technologies Ethical,legal,and social implica-tions(as

154、pects)with the widespread use of AI(ELSI/ELSA)Massive or Ultra-large Systems of systems(SoS)Self-configurable integration and interoperability Self-Adaptive AI and Massive use of Reinforce-ment Learning General Purpose Artificial Intelligence Hyperconvergent Infrastructure Autonomy forlow-cost,attri

155、table weapons Offensive cyber Multi-use autonomous defences and how does it change theair power game Climate change How do our systems work in a changing environment Applications ofquantum technology Interruption of the hypersonic kill chain Space denial/protection Interoperability(Standards)managin

156、g operations in the EW spectrum Use and democratisation of direct energy weapons(DEW)AI Automation Hyperspace AI/Automation within cyber Human trust in autonomySET:Cognition and AI for RF-Sensor Technology RF Technology for counter-UAS applications RFand digital technology Space Situational Awarenes

157、s Quantum Algorithms for Data Fusion IntegratedPhotonic Sensing Neuromorphic Processing Crystal Vacancy Centre Sensors AI for underwater autonomyexploration/sensing Certification of AI systems Metal-Organic Frameworks Neuromorphic sensing Mid-IR lasers Quantum lidar Photonic integrated circuits Meta

158、lenses Single photon counting detectorarrays in the infrared MSG:Digital TwinEDT Survey of STO PanelsThe STO network consists of approximately five thousand active science workers.A survey was preparedand distributed to all STO panels and members engaged in the research network to leverage this expe

159、rtise.The survey was available in 2022 from early September to the beginning of November.It collectedinformation on technology readiness levels,maturity forecasts and critical keywords associated with theEDTs and TFA.The survey results have provided a wisdom of the crowd assessment,with individualsc

160、ontributing appraisals only to those EDTs for which they have experience or interest.Out of approximately 5000 active participants within the STO network,approximately 8%respondedto the survey.Such a response rate is considered average for an external survey of this type.Figure 2.5demonstrates the a

161、ssessments by EDT.2.3.4ReportsOver the last two years,the STO has responded to NATO leaders challenging the NATO communityto improve awareness and exploitation of EDTs.In response,the STO developed a comprehensiveassessment of EDTs,and their accompanying S&T ecosystem 2.Since then,the STB,through th

162、eOffice of Chief Scientist,has supplemented its earlier EDT studies,responding to NATOs ImplementationStrategy for EDTs.This resulted in a series of NATO Chief Scientist reports and NATO documents(AC/323)summarising or exploring the S&T landscape.Those publicly released or unclassified reportsmay be

163、 found at https:/www.sto.nato.int/Pages/NATO-Chief-Scientist-Reports.aspx and include studieson women in the armed forces,Quantum technologies,AI,Autonomy,human factors for special forces,CBRN,third country S&T developments and technology weak signals.In addition,STO reports such asthose on biotechn

164、ologies 34 and CBRN 35 have provided unique and valuable insights into the futuredevelopment of associated S&T areas.2.3.5AnalyticsTo provide a solid analytic basis for the assessment of EDTs,the NATO Office of the Chief Scientistco-developed with the NATO Communications and Information Agency(NCIA)

165、the S&T EcosystemAssessment Model(STEAM)Figure 2.6.STEAM is hosted by the NATO software factory and is anAI-enabled analysis tool that assesses various published reports,journal articles and pre-prints to explorecurrent trends,collaboration,national focus,and research interests.The system pulls from

166、 over 7 milliondocuments and 200 million abstracts,providing a representative sampling of current research activitiesover the last five years.The underlying data is pulled from Microsoft Academic 36,arXiv 37(Physics,2.3 Information Sources16Figure 2.4:Microsoft Forms Survey of STO network on S&T tre

167、nds.17Chapter 2.MethodologyEDT Survey ResponseFigure 2.5:Responses by EDT from the STO EDT Survey.Mathematics,Statistics,Computer Science,Electrical Engineering,Economics,Quantitative Biology,Quantitative Finance),MedRxiv 38(Medicine)and bioRxiv 39(Life Sciences).In addition,theintegration of new so

168、urces(such as patent data)and reports(e.g.PLOS(the Public Library of Science)40)is being evaluated for use in future analyses.Nevertheless,STEAM is still a work in progress,withits analysis more indicative than definitive,especially given the lack of non-English language sources.NCI Agency|Demonstra

169、tion:Weak Signals in Technology TrendsTime Period:10 YearsUse Case 1:Analysis by Location(China)In a fixed time period(2018-2021)Publications by categoryover timeTop categories by number of publicationTop Affiliations in selected location perselected time periodUse Case 2:Analysis by Affiliation(Sta

170、nford University)In a fixed period of time(2018-2021)Affiliations Network GraphTop Co-Affiliations by CategoryTop Co-Affiliations per a selected category of interest(ex:Computer Science)LocationInteractions per selected time periodCorpus:1,5 Million of scientific papersAffiliations:8364Locations:156

171、Data Sources:1.ArXiv2.Microsoft Academic GraphFigure 2.6:Early Demonstration of S&T Ecosystem Assessment Model(STEAM)Use Cases.2.4 Alliance and Partner Research Programmes182.4Alliance and Partner Research ProgrammesMost STO-sponsored S&T activities are undertaken through national collaborative rese

172、arch activities.The STO maintains cross-alliance visibility on national research activities and priorities.These activitiesprovided insights into emerging technology areas and activities of importance to the Alliance.In particular,the following research programs and S&T strategies were of considerab

173、le value in understanding nationalstretch objectives in defence and security S&T:Australia 41,US(DARPA)42,UK 43 and France44.In addition,several countries have developed lists of emerging and disruptive technologies.Tworecent US lists of particular value were 14 and 15.The first list provides the fo

174、llowing breakdown andreflects a whole-of-nation perspective:Advanced Computing:Supercomputing Edge computing Cloud computing Data storage Computingarchitectures Data processing and analysis techniquesAdvanced Engineering Materials:Materials by design and material genomics Materials with newpropertie

175、s Materials with substantial improvements to existing properties Material property characterisationand lifecycle assessmentAdvanced Gas Turbine Engine Technologies:Aerospace,maritime,and industrial development andproduction technologies Full-authority digital engine control,hot-section manufacturing

176、,and associatedtechnologiesAdvanced Manufacturing:Additive manufacturing Clean,sustainable manufacturing Smart manufac-turing NanomanufacturingAdvanced and Networked Sensing and Signature Management:Payloads,sensors,and instruments Sensor processing and data fusion Adaptive optics Remote sensing of

177、the Earth Signature manage-ment Nuclear materials detection and characterisation Chemical weapons detection and characterisation Biological weapons detection and characterisation Emerging pathogens detection and characterisation Transportation-sector sensing Security-sector sensing Health-sector sen

178、sing Energy-sector sensing Building-sector sensing Environmental-sector sensingAdvanced Nuclear Energy Technologies Nuclear energy systems Fusion energy Space nuclear powerand propulsion systemsArtificial Intelligence:Machine learning Deep learning Reinforcement learning Sensory perceptionand recogn

179、ition Next-generation AI Planning,reasoning,and decision making Safe and secure AI Autonomous Systems and Robotics:Surfaces Air Maritime SpaceBiotechnologies:Nucleic acid and protein synthesis Genome and protein engineering,including designtools Multi-omics and other biometrology,bioinformatics,pred

180、ictive modelling,and analytical tools forfunctional phenotypes Engineering of multicellular systems Engineering of viral and viral delivery systems Biomanufacturing and bioprocessing technologiesCommunicationandNetworkingTechnologies:Radio-frequency(RF)andmixed-signalcircuits,antennas,filters,and co

181、mponents Spectrum management technologies Next-generation wireless networks,including5G and 6G Optical links and fibre technologies Terrestrial/undersea cables Satellite-based communications Hardware,firmware,and software Communications and network security Mesh networks/infrastructureindependent co

182、mmunication technologies Directed Energy:Lasers High-power microwaves Particle beamsFinancial Technologies:Distributed ledger technologies Digital assets Digital payment technologies Digital identity infrastructureHuman-Machine Interfaces:Augmented reality Virtual reality Brain-computer interfaces H

183、uman-machine teamingHypersonics:Propulsion Aerodynamics and control Materials Detection,tracking,and characterisation defence Networked Sensors and Sensing:Not available19Chapter 2.MethodologyQuantum Information Technologies:Quantum computing Materials,isotopes,and fabrication techniquesfor quantum

184、devices Post-quantum cryptography Quantum sensing Quantum networkingRenewable Energy Generation and Storage:Renewable generation Renewable and sustainable fuels Energy storage Electric and hybrid engines Batteries Grid integration technologies Energy-efficiencytechnologiesSemiconductors and Microele

185、ctronics:Design and electronic design automation tools Manufacturingprocess technologies and manufacturing equipment Beyond complementary metal-oxide-semiconductor(CMOS)technology Heterogeneous integration and advanced packaging Specialised/tailored hardwarecomponents for artificial intelligence,nat

186、ural and hostile radiation environments,RF and optical components,high-power devices,and other critical applications Novel materials for advanced microelectronics Wide-bandgap and ultra-wide-bandgap technologies for power management,distribution,and transmissionSpace Technologies and Systems:On-orbi

187、t servicing,assembly,and manufacturing Commoditisedsatellite buses Low-cost launch vehicles Sensors for local and wide-field imaging Space propulsion Resilient positioning,navigation,and timing(PNT)Cryogenic fluid management Entry,descent,andlandingThe second list provides a more defence-oriented pe

188、rspective:Seed Areas of Emerging Opportunity:Biotechnology Quantum Science Future Generation WirelessTechnology(FutureG)Advanced MaterialsEffective Adoption Areas:Trusted AI and Autonomy Integrated Network Systems-of-Systems SpaceTechnologies Renewable Energy Generation and Storage Advanced Computin

189、g and Software Human-Machine Interfaces Defence-Specific Areas:Directed Energy Hypersonics Integrated Sensing and CyberSimilarly,the European Union conducts technology trends assessments through the European DefenceAgency and European Commission.In particular,the European Commission(EC)Joint Researc

190、h Center(JRC)annually evaluates emerging and disruptive technologies,determined through workshops andanalysis of publications and patent data.The latest in these reports may be found in 16.Medicine&Biotechnology:BiSCaOantisepticsCELMoDsforcancerChemodynamictherapyCRISPR-based diagnostics Direct RNA

191、sequencing DIY artificial pancreas eCIRP Facial genotyping FAPIPET/CT Firibastat Flash radiotherapy Flualprzolam Heritable human genome editing Inebilizumab Intravascular Lithotripsy Mavacamten Molnupiravir Nanocatalyst cancer therapy Soluble immunecheckpoints Spatial and temporal omics UrobiomeEngi

192、neering&Physics:Atomic-scale manufacturing Autonomous surface ship Gas chromatography-ionmobility Nanoparticle exosolution Optoelectronic synapses Passive radiative cooling Rotating detonationrocket engine Space-air-ground integrated network Spintronic terahertz emitter TwistronicsMaterials:Eutectog

193、el High entropy carbides Hybrid nanofluid Ionic skin Kagome superconductors Martian concrete Monolayer MA2Z4 Monolayer tellurne Mycelium-based materials NB2CTx mxene Nickelate superconductor Orgonohydrogel Quantum metasurfaces Twin graphene Twisted bilayergraphene Ultrawide bandgap semiconductor V2C

194、Tx MxeneEnergy:Aqueous Al ion batteries Calcium batteries Dendrite-free Zn battery Hydrovoltaics Mg metalbatteries Multi-energy microgrids Potassium metal batteries Sodium CO2 batteries Zinc CO2 batteries Zinc-organic batteries Zinc ion capacitorICT-Information and Communication Technologies:3D mult

195、i-object tracking Cognitive digital twin Directed Acyclic Graph blockchain DNS-over-HTTPS Dual blockchain Wearable computing Edgeartificial intelligence External Human-Machine Interfaces Few-shot object detection Fog robotics Healthchatbot Human digital twin Internet of space things Programmable wir

196、eless environments Unpairedimage translation Versatile video coding Wireless time-sensitive networkAgriculture&Environment:Agrophotovoltaics Biomass chemical looping gasification Interfacial solarevaporators Microplastics removal Nanoagriculture Nanofertilisers Plasma agriculture Plastic chemicalupc

197、ycling2.4 Alliance and Partner Research Programmes20Societal Issues:AI ethics Fake news detection Green AI Misinfodemics Peer-to-peer energy trading Real-world clinical outcomeFurther,the European Defence Agency(EDA)has a structured approach to developing research areasthrough the Capability Technol

198、ogy Areas(CapTechs)and the Overarching Strategic Research Agenda(ORSA)Technology Building Blocks(TBB)45,46.A slightly modified,expanded and consolidatedversion is shown below:Air Autonomous Air Vehicle Operation Cooperative Air Vehicle Operation Detect,Sense and AvoidSystems System Diagnostics,Fault

199、 Prognostics,and Self Repair Human-Machine-Interface and CognitiveErgonomics Propulsion,Power Generation and Distribution Secure Command and Control Systems Rotorcraft Next Generation High-Performance Vertical Lift Fixed WingMissiles and Munitions Munition Life Management Insensitive Munitions Fuzin

200、g and Ignition Systems Precision Guided Munition and Missiles High-Performance Gun Launch and missile propulsion Railgunweapon Improved tools and methods for qualification and safety IED Detection and Defeat Technologies High-performance,low-sensitivity and REACH-compliant energetic materials Improv

201、ed Warhead andPenetrator design Pyrotechnics for Decoying and Obscuring New Production Techniques for MunitionComponents Education and Training for Ammunition TechnologiesCBRN&HF Human Autonomy Teaming Human Performance Monitoring&Enhancement CustomisedTraining Integration of human clothing and equi

202、pment in platforms Personal Protective Equipment Detection,Identification and Monitoring(DIM)of CBR CBRN Hazard Management CBRN modellingand simulation Protection of Critical Infrastructure(PCI)from CBRN Protection of food and water supplyfrom CBRN Assessment,Diagnosis,and Medical Countermeasures of

203、 CBR Hazards Human Resources&Social SciencesComponents RF Photonics IR imaging Detector&sources Terahertz detectors&sources MicrowavePower RF Transceiver modules Enabling Components for Advanced Antennas ADC&DAC SignalGeneration and Time Reference System-on-Chip High voltage SiC devices and related

204、energy storagefor pulsed power applications Defence critical technologies supply chain Advanced Packaging,PCB andThermal Management TechnologiesCyber R&T Cyber Defence Situation Awareness Cognitive Science with cyber implications Convergencebetween Cyberoperations and Electronic Warfare Cross-cuttin

205、g cyber defence for land,maritime,air andspace The protection of military communications and information systems Quantum computing andcryptography with cyber implications Autonomous cyber response capabilities Modelling and Simulationfor Cyber defenceEnergy Alternative fuels and drive/propulsion sys

206、tems Energy storage:electrical,electrochemical,mechanical,structural and thermal Engine and power distribution system efficiency technologies Energymanagement technologies:innovative and efficient systems Solar energy generation(thermal and electricalgeneration)Militarisation of environmental techno

207、logies:water,wastewater and energy from waste Energyharvesting/scavenging Wind energy Energy and environment technology systems integrationInformation Management and Processing Information from Heterogeneous Sources Coalition NetworkSecurity and Protection and Interaction with commercial technologie

208、s Cognitive Radio Tactical CloudInfrastructure for C4ISR Systems Electronic Support Measures(ESM)for Communications Internetof Things(IoT)for Defence Information Process Enhancement by using AI and Big Data Long RangeCommunication Software Defined Networking(SDN)Land Land Systems Architecture&Integr

209、ation Power generation,storage and management for LandSystems Passive and active protection for land Systems Less-than-lethal effectors Manned/unmannedteaming,adaptive cooperation between manned and unmanned systems with different levels of autonomy Target/Threat recognition and identification Healt

210、h and Usage Monitoring Novel User Interfaces forSoldier Assets integration/control Mobility and Counter-Mobility Weapon system integrationMaterials Light Weight for High-Performance Structures Materials,structures for Protection Against Mil-itary Threats High-temperature materials Camouflage and sig

211、nature management technologies Emergingmaterials for future platforms Materials,structures&concepts for platform monitoring New manufacturing,joining and repair processes Surface engineering for maximum lifetime and hostile environments Advancedand smart textiles for soldier systems and platforms Co

212、mputational Design and materials modelling21Chapter 2.MethodologyMaritime Communication and distributed sensor networks,surface and underwater Simulation andTrainingPlatformSurvivabilityandOperabilityinchallengingconditionsEnergyandPropulsionIncreasedAutonomy and Robotics Identifying and Countering

213、Threats,High-Energy Weapons Integration SmartIndustrialisation and Predictive MaintenanceNavigation Navigation in GNSS denied environment Position,Navigation and Timing(PNT)superiorityand integration in operations and systems Guidance and control in challenging environments Autonomousand automated G

214、NC and decision-making techniques for manned and unmanned systems Fault tolerancecontrol(FTC)Multi-Robot control and cooperation Precision guidance and control of weapons Keyenabling and performance enhancing GNC technologiesOptronics Hyperspectral/multispectral Imaging System Passive Imaging System

215、s Novel optical config-uration Active Imaging System Image enhancement Image processing EO Countermeasure systems Laser Weapon System Modelling and simulation Data fusion and system integrationRadar Detection,Tracking and Recognition of Challenging Targets Advanced front/back end ElectronicWarfare M

216、ulti-Platform RF Systems Cognitive Radars Scalable Multi-function RF Sensors(SMRF)Specific Radar Applications Disruptive Concepts Common EU Benchmarks for Validation,Verificationand StandardisationSimulation Integrated Live,Virtual and Constructive(I-LVC)for Training,Simulation and SeriousGames Solu

217、tions Artificial Intelligence(AI)and Big Data(BD)for Decision-Making Support Immersive,Virtual and Augmented Reality Cyber Defence Simulation Joint Strategic,Operational and Tactical levelsimulators Modelling&Simulation as a Service(MSaaS)for the synthetic environment and rapid scenariogeneration Si

218、mulation for Systems of Systems(S3)Recognised Space Picture(RSP)Defence SatelliteReconnaissance SystemsSpace Architecture&Policy Earth Observation&Reconnaissance Space Situational Awareness(SSA)&Space Operations Positioning,Navigation and Timing(PNT)Satellite Communications Quantum andother Disrupti

219、ve Technologies Data&Signal Processing for Decision-Making Launch,Space&GroundSegment Technologies Protection2.5MetaStudies,Articles and ReportsMany nations,academic and commercial interests are involved in research or developing reports on future(science&)technology trends or strategic foresight.Th

220、is report draws on a diverse collection of journals,newspapers,magazines and web sources identified and consolidated through various news aggregatorsand independent research.Major foresight studies also provide a core foundation for this report.Ofparticular use in the drafting of this report were th

221、e 2022 Gartner Hype Cycle for Emerging Technologies24,Deloitte Insights-2022 Tech Trends 47,Future Today Institute Tech Trends 2022 48,OECDScience,Technology and Innovation Outlook 2021 49,Imperial TechForesight 50,The Report ofthe Commission on the Geopolitical Impacts of New Technologies and Data(

222、North Atlantic Council-Geotech Center)51,Articles from the NATO Defence College(e.g.52,53,54,55),CongressionalResearch Service Reports such as 56,67,76,80,81,82,83,84,85,57,58,59,60,61,62,63,64,65,66,63,68,69,70,71,72,73,74,74,75,77,77,78,79,Federation of American Scientists 86,EuropeanParliament 87

223、,RAND reports 88,89,90,UNESCO Science Report 91,World Economic Forum92,Brookings 27,and summaries of international technology forecasts found in 93.For AI and thedevelopment of STEAM,the Center for Security and Emerging Technologies(CSET)94 at GeorgetownUniversitys Walsh School of Foreign Service wa

224、s an especially valued source of information and insights.3.EDT Analysis3.1DescriptionThis volume has separate appendices that provide a more detailed exploration of each EDT,drawingheavily upon STO research and technology watch activities,quantitative analysis,surveys,serious gamesand expert opinio

225、n.This section also includes Conjecture Cards,short vignettes that describe the potentialfuture application or potential consequences of these technologies.They are included to help contextualisethe potential impact of these technologies.Each appendix follows a set structure:Title:A short title for

226、the EDT.Quote:These provide context for the discussion.Definition:A useful definition of the EDT in question.Where appropriate,these are drawn fromapproved NATO or national sources,although they may be expanded as necessary to describe theEDT in question.Keywords:Useful keywords that may aid indepen

227、dent topic exploration.Overview:An expansive overview of the EDT in question.It is broken into sub-sections represen-tative of the technical focus areas(TFA),or if you prefer,a second level representing importantsub-aspects of the EDT.Military Implications:A consideration of how this EDT will impact

228、 NATOs(e.g.BLUE)operational and enterprise capabilities(Enable,Prepare,Project,Engage,Protect,Sustain,Inform,C3).This is followed by discussing the unique implications for forces or groups(e.g.RED)thatmay threaten NATO and Alliance nations.Interoperability:Discusses the EDTs interoperability and sta

229、ndardisation challenges.S&T Development:Explores the implications of developments in this EDT on the other EDTs(e.g.Data as it enables developments in AI).STO Activities:A sampling of current STO activities relevant to this EDT.STEAM:Publication Keywords.Given the shear volume of charts created a se

230、parate limiteddistribution report will be developed at a later date covering all EDTs and TFAs showing:keywords,23Chapter 3.EDT Analysisnational technology leadership,publication trends,institutional leadership and institutional collab-oration.Appendix L provides an example of these charts and Volum

231、e 1 of this report provides asub-set for EDTs only.Survey:Simplified Table of Impact,TRL and Horizon.Appendix K provides a comprehensiveassessment of the survey results showing:Keywords(Word Cloud),Impact Assessment,TRLAssessment and Horizon Assessment.Conjecture Cards:Representing small vignettes d

232、esigned to stimulate discussion or support thedevelopment of a technology futures(serious)game.3.2Exploration of Selected EDTsAs of 2022,there are nine canonical EDTs and one additional for the purposes of this report.These are inalphabetical order:AI:Artificial Intelligence(Appendix A).Autonomy:Rob

233、otics and Autonomous Systems(Appendix B).BHET:Biotechnology and Human Enhancement(Appendix C).Data:Big Data and Information Communication Technologies(Appendix D).E&EM:Electronics and Electromagnetics(new supplemental EDT)(Appendix E).Energy:Energy and Propulsion(Appendix F).Hypersonics:Hypersonic T

234、echnologies(Appendix G).Materials:Novel Materials and Advanced Manufacturing(Appendix H).Quantum:Quantum Technologies(Appendix I).Space:Space Technologies(Appendix J).Two additional appendices follow these:STO Experts Survey:An analysis of the STO-wide survey on the state,rate,and impact of EDTs(App

235、endix K).S&T Analysis Model(STEAM)Results:A sample analysis of the EDTs using the S&T EcosystemAnalysis Model(Appendix L).4.Conclusion“If you make a great number of predictions,the ones that were wrong will soon be forgotten,and theones that turn out to be true will make you famous.”-Malcolm Gladwel

236、l 95Technology and SocietyMastering the R&D,concept development,and military application of EDTs will expand NATOsability to respond at an enterprise level and support multi-domain operations.However,the predictionof militarily relevant S&T developments over a twenty-year period is non-trivial.To do

237、 so,the STO hasdrawn from various data sources,analytical methods,and expert opinions to outline a realistic estimate ofthe state,rate of development,and impact of EDTs.Ultimately,this report has outlined how EDTs will disrupt,degrade,and enable NATO militarycapabilities in the 2023-2043 timeframe.U

238、nderstanding why they present a problem or opportunity,howthey are expected to manifest,and what this will mean to the Alliance is an excellent first step and willhelp ensure NATO remains technologically prepared and operationally relevant.AppendicesA.Artificial IntelligenceWe have to improve scient

239、ific understanding amongst AI paradigms so as to build AI that benefitshumanity and the planet.The world will need a principled AI education for all,since AI will be the keytechnology of the next decades,if not the XXI century.-Luis Lamb 96Artificial IntelligenceDefinitionAI refers to systems that d

240、isplay intelligent behaviour by analysing their environment and taking actionswith some degree of autonomy to achieve specific goals.AI-based systems can be purely software-based,acting in the virtual world(e.g.voice assistants,image analysis software,search engines,speechand face recognition system

241、s).Alternatively,AI can be embedded in hardware devices(e.g.advancedrobots,autonomous cars,drones or Internet of Things applications)97.Artificial Intelligence(AI)KeywordsArtificial Intelligence(AI)Machine Intelligence Deep Learning(DL)Machine Learning(ML)Artificial Neural Networks(ANNs)Machine Lear

242、ning(ML)Expert Systems Semantic Analysis Supervised Learning Unsupervised Learning Reinforcement Learning Clustering Deep Fakes Machine Vision Chat-bot Decision Trees Data Science Genetic Algorithm Autonomy ArtificialGeneral Intelligence(AGI)Responsible Artificial Intelligence(RAI)Generative Adversa

243、rial Networks(GANs)Explainable AI(XAI)Observable AI Trust Trustworthiness AI Robustness Resilience Generative Adversarial Network(GAN)DevelopmentsOverviewSince the publication of the last NATO S&T trends report 2,Artificial Intelligence(AI)has evolvedrapidly and in surprising directions across multi

244、ple domains 98.At the same time,AIs wide and imagi-native application has created a climate of increased expectations regarding its possible implementation.27Chapter A.Artificial IntelligenceThe intrinsic value of AI and machine learning is centred on its predictive capabilities.However,to extractth

245、ese technologies full potential,organisations must restructure and evolve 98.It is worth mentioning that currently,AI presents some technical inefficiencies,limitations,andhuman-trust issues concerning its underlying methodology,which are becoming more obvious due toits increasing use.AIs large and

246、diverse applications are game-changers that must still be enabled byappropriate methodologies(e.g.AI methods).The growing need to reduce disparities is a weak technologysignal that indicates how the application of divergence methods will potentially lead to the next AI winterin 5-10 years 2.For this

247、 appendix,AI is understood to encompass four main areas of R&D activity,which are seen to bedisruptive:(1)AI Methods;(2)AI Application;(3)M-M Symbiosis;and,(4)Resilience.Breakthroughsin these areas will correct and reduce divergences,drive increased adoption,and enhance the robustness,utilization,an

248、d reliance on AI across the Alliance and with potential adversaries.Expert knowledge or criteria&logical reasoningThe first wave of AI is represented by expert knowledge or other authoritative sources and encoded into a computer program in the form of an expert system.Examples:Resource Allocation,Ma

249、intenance,Inventory ControlMachine/Statistical LearningThe second wave of AI is based on machine learning(ML),or statistical learning,and includes voice recognition,natural language processing,computer vision etc.Examples:Face Recognition,Spam filters,Medical Diagnosis,Semantic AnalysisContextual Ad

250、aptionThird wave AI technology combines the strengths of first-and second-wave AI,and is also capable of contextual sophistication,abstraction and explanation.Examples:Autonomous Vehicles,Cyber Agents.SECOND WAVEPerceivingLearningAbstractingReasoningPerceivingLearningAbstractingReasoningPerceivingLe

251、arningAbstractingReasoningRelative Level of CapabilityFIRST WAVETHIRD WAVEFigure A.1:The Three Cycles of AI(CREDIT:Adapted from 99).The AI was conceptually envisaged before 1900 99,and it has undergone three main cycles oftechnological and scientific development 100(See:A.1):Expertor(handcrafted)kno

252、wledge-basedAI:ThisfirstcyclecompilesAIsembryonicknowledge-based state grounded in experts rules-based decision-making approaches(e.g.hard-coded if-thenstatements,decision trees,Booleanand fuzzylogic),which constrain systems reasoning andlearningabilities.Data-based AI(statistical Learning):The seco

253、nd cycle advances AI systems by developing andapplying statistical methods trained on(pre-labelled)data for specific purposes.Those methodsincludeMachineLearning(ML)andDeepLearning(DL),whichofferawidearrayoftechniquesandintroduce logic-senses and(auto-)learning abilities(e.g.supervised,unsupervised

254、and reinforcementLearning).These second-wave systems rely heavily on the data they have been fed and trained on.28New or next-generation of AI(contextual adaptation):Currently,the third wave(or Neurosym-bolic 101)AI presents considerable developments in sensing and perception abilities enablingconte

255、xtual systems adaptation.It has been characterised by ubiquitous intelligent(or autonomous)systems(also called embedded-AI),bio-,and human-inspired learning methods(e.g.artificialneural networks such as deep and machine Learning)99.Figure A.2:Neurosymbolic AINowadays,the development of AI is still c

256、har-acterised by hype and bluster 100.AI technol-ogy is considered a scientific area of developmentlargely attracting national attention;and a(general-purposes)technology with extensive dual-use formilitary and civilian applications 86,102,103.Its wide array of applications spans from cogni-tive aug

257、mentation to multi-domain operations plat-forms,continuingwithembodiedAI(robotics)andadvanced algorithms.AI offers an unprecedentedopportunity to strengthen NATOs technologicaledge and will support the Alliance in fulfilling its three core tasks 104.However,increased reliance onAI capabilities will

258、also expand the threats the Alliance will face.Over the next 20 years,AI will continue to play a significant disruptive role through its impact onseveral areas 2,such as:Exploitation of(very)large data sets,including publicly available data for system training anddevelopment;Widespread deployment an

259、d use in cyber-physical systems;Novel areas of application,driven by greater investment in and wider adoption of AI techniques;Decision making and optimal control(e.g.speed,power systems,investment,etc.);Computation,such as advances in everywhere/edge computing,ubiquitous sensors,database design,dev

260、elopmental tools,cloud computing,new algorithmic approaches and using AI to bootstrap thedevelopment of AI;and,Development of advanced large data analysis tools and computer vision.However,the limitations of AI become more obvious(Figure A.3)as its deployment expands.Brittle-ness jeopardises AI syst

261、ems performance and reliability in simulation and training(e.g.different modelsettings or data inputs,uncertain or new scenarios).Unpredictable AI systems are further exposed todata poisoning and adversarial attacks leading to systems failures and malfunctions 105.Moreover,the rationale behind any A

262、I/ML recommendations can be opaque(e.g.AI Black-box 106).Thislack of transparency harms the development of trust in Human-Machine(H-M)symbiosis and teamingrelationships.Stressing the need to revisit the mathematical,algorithmic,and technical foundations ofthe enabling science 105.The lack of acknowl

263、edgement of the pitfalls is usually characterised byunintended,intelligence or automation bias 107,leading military and political decision-makers andoperators to unfounded-misplaced trust,mistrusting and over-relying on flawed AI systems.In addition,the lack of transparency reduces the tractability

264、of data sources and can ultimately impact attack detection105.Thus leading to concerns about the malicious use of AI and the return to an AI winter 2.On the other hand,considering the global attention that AI is getting in both the private and publicsectors,AI has been compared to the introduction a

265、nd proliferation of electricity as a transformativetechnology 103,108.Similar to nuclear weapons proliferation in history,as argued by 109,this couldraise security concerns as the Sino-US competition escalates,potentially shifting global powers.Opposing views 107 argue that the fierce competition ov

266、er AI does not constitute a security dilemma.Still,policymakers should mind the technological and political risks associated with AI and its competition.29Chapter A.Artificial IntelligenceFigure A.3:Understanding AIs limitations and drawbacks(CREDIT:105.Lower right icon created byRanah Pixel Studio)

267、.Some of these risks include the Sino advance,diffusion and dissemination of Chinese technology standardsin AI systems and the global market.Hence,posing a security concern and opening the door to technologystandard competition,patent protection,lawfare and(design)policy writing.Consequently,stressi

268、ng theneed for checks and balances towards a more democratic AI,ethical,legal and moral(ELM)use andresponsible use of AI(RAI).This calls for international(military)agencies,DoDs and governments tocooperate and regulate AI,or at least to define a shareable ethical vision and responsible criteria for

269、its use1101.To mention a few attempts,the European Data Act(as the first deliverable under the EuropeanStrategy for Data)and AI Strategy,NATO Summary AI Strategy and its Principles of Responsible Use(PRU)104,111,112,113.Much has been written on the growth of AI and its importance for the fourth indu

270、strial revolution103,114,115,as AI is a critical enabler for other scientific and technology areas 103-see Figure A.5.In addition,new developments and leaps concentrate on compiling current knowledge and advancing newlearning techniques while enhancing the robustness,generalisation and trustworthine

271、ss of AI systems101.Furthermore,research on advancing Man-Machine Symbiosis(or Human-Machine Teaming),explainable AI(XAI),hybrid AI techniques,and advanced training methods are at the forefront in thedefence and security sphere.Advanced AIR&D in this area began with early narrowed AI expert systems

272、compiling rules-based approaches,oftenconfined by human experts knowledge of mathematical and statistical models(e.g.linear algebra,linearand multiple regression).Machine-Learning(ML),natural language processing(NLP)and ComputerVision in AI-based systems,currently applied and known as weak or narrow

273、 AI,soon complementedthese developments.Further improvements in ML and learning infrastructures lead to the second wave ofAI developments.Advanced algorithms perform on a learning spectrum(supervised,semi-supervised and unsupervisedlearning,as well as transfer learning)based on the Algorithms abilit

274、y to label and cluster data identifying1The OECD offers a live comprehensive repository and update of AI-related policy information,OECDs strategies andpolicies from 44 countries and the EU(https:/oecd.ai/en/dashboards/overview)30inner-data patterns.Thus,enhancing perception while introducing the ab

275、ility to learn and infers contextat the cost of reasoning abilities.Downturns are addressed in the third cycle of AI,also called nextgeneration,combining elements of both prior waves towards general artificial intelligence(GAI).Critical areas of research and key points are:Linear Algebra 116 Linear

276、Regression Logistic Regression Artificial Neural Network(ANN)ConvultionalNeuralNetwork(CNN)(Residual Networks,Conv1x1,Inception,UNet Dropout,Batch Normalization,GlobalPooling,Auto Encoder,Generative Adversar-ial Network(GAN),Transfer Learning)117,118,119,120Hierarchical Recurrent Neural Network(HRNN

277、)(TanH,Sigmoid,LSTM,GRU,Bidi-rectional Recurrent Neural Network(RNN),Seq2Seq,Mechanism of attention,Trans-formers)119,121Deep Learning(DL)(hybrid architectures,Fractal Networks)Machine Learning Algorithms(NaiveBayes,SVM,KNN,Decision Tree,Gradi-ent Boosting,XGBoost,CATBoost,GenericAlgorithms)122Reinf

278、orcement Learning(RL)(Q-Learning,Deep-Q-Learning,Federated Learning,Ma-chine Common Sense,Learning with Re-duced Labelling,OpenAI Gym)42,119,122,123,124Quaternion Valued Neural Network 119,125Fundamentally,AI progress aims to expand its mathematical,computational,conceptual,and algo-rithmic underpin

279、nings.Advancements are driven by a reduced need for data,enhanced agility,increasedresilience,improved problem-solving,and generalisation of scenarios while synergistically cooperatingwith human elements or wetware(i.e.,biological)systems.Figure A.4:Artificial Neural Network-AI Methods(CREDIT:WikiCo

280、mmons)AdvancedAlgorithms(AA)isanalysedalongtwo dimensions:learning architectures and devel-opments in AI fundamentals.However,AIs nat-ural limitations and drawbacks heavily constrainboth dimensions(see Figure A.3).Nonetheless,they attempt to address intrinsic issues such asproblem set optimisation,t

281、ransparency,and trustwhile facing brittleness and generalisation issues.Overcoming these challenges will increase mili-tary decision-making effectiveness and efficiencythrough symbolic,and sub-symbolic approaches126,neural networks,and learning architectures.On the other hand,AA methods are evolving

282、 towards next-generation AI methods that build onbiological modelling and mimic human cognitive functions.Examples of such approaches are:Long short-term memory neural networks 119,121,127;Hierarchical recurrent neural networks 119,121;Residual neural networks 119,128;Convolutional neural networks 1

283、18(with significant research interest on convolutional autoen-coders 117,and dilated convolutional neural networks 119);and Quaternion-valued 119,125.31Chapter A.Artificial IntelligenceThe above AA is integrated within Learning architectures,making AI more agile and resilient whileusing fewer data 4

284、2,129.Different methods such as Deep-Q learning 119,122,federated learning123,130,and machine common sense 42,124 are promising research areas,though antagoniststrategies involving generative adversarial networks(GAN)119,120.Siamese networks 119,131 haveproven themselves especially powerful.AI Appli

285、cationFigure A.5:Can AI be an inventor?The application of AI is considered a priority S&Tinvestment area and is currently being driven byindustry 86,132.On the one hand,advance-ments in the application of AI in the public do-main have been significantly enabled by the pro-duction and availability of

286、 open-source tools anddata(e.g.Enron dataset 133).On the other hand,military AI applications are usually limited to con-text.They lack specific datasets for training algo-rithms,though Alliance nations are still investingand progressing on AI applications 107,134.AIcontinues to be embedded within ma

287、ny systems,processes,and defence capabilities.Some of the more exciting and potentially disruptive applications arepresented below 107,134,135:Decision making 136,137,138,139,140 Cybersecurity Automated Vehicles 135 Artificial Social Intelligence Cognitive Warfare 141 Automating Disinformation Geosp

288、atial AI High-Resolution Image Recognition Semiconductor DesignGeneration and Identification of Deep-fakes Data-Driven Behaviour Modelling Dynamic Airspace Control Scientific and Technological Discovery Text to Image Synthesis Human Terrain NavigationFigure A.6:AI and Decision Making CREDIT:140The e

289、xtent to which AI enables society and scientific domains is still being determined.However,32AIs unique ability to unlock or act as an enabler of new capabilities and scientific discoveries acrossmultiple disciplines(e.g.scientific,mathematical and engineering,educational 142,medical 143,144,and oth

290、ers)contribute to its comparison to other past revolutionary technologies such as electricity 103.The National Security Commission on Artificial Intelligence(NSCAI)contextualised AI as theengine of invention 103 and proposed some examples(Figure A.7),which have been complemented inthe sections below

291、.Instances of the impactful nature of AI span from NLP,evolve into more humanisedNLU(e.g.language translation)and continue towards AI-enhanced modelling and simulation based onintegrated multi-modal cloud infrastructures(e.g.transforming crises responses).Alternatively,combinedwith autonomous techno

292、logies such as drones,AI can empower first respondent forces by augmentingtheir ability to orient themselves and map out unknown surroundings.Lastly,among other examples,AI contributes to operational and strategic thinking and planning while accelerating C2 and sharpeninglogistics.Figure A.7:AI as t

293、he Engine of Invention(CREDIT:103).AA methods such as ML and DL contribute to increasingly diverse data-driven decision-making andintelligence analysis application 145.In particular,AI-enhanced human decision-making is used forlarge data collection from multiple sources,and anomaly detection within

294、a system 135.The speed of decision-making processes(e.g.the observe-orient-decide-act(OODA)loop)providescritical leverage to gain an advantage,and decision-making superiority over adversaries 145,103.The extended use of AI-enabled techniques exposes and empowers cybersecurity.AA,automatedand autonom

295、ous systems paired with human expert systems enable a predictive model of cyber-attacksand support identifying and patching vulnerabilities;or detecting and defending against threats 135,145.33Chapter A.Artificial IntelligenceSome of these threats include using AI-enabled systems in the context of C

296、ognitive warfare.Indeed,cyber hacking,technological espionage,blackmail,and illicit technology transfer 103 are used as meansof hybrid warfare and mis-/dis-information techniques 103,145,146,147,148.The previous threatshighlight the potential role of AI/ML regarding improvements in deception and inf

297、ormation operations,automating disinformation,and in the generation and identification of deep-fakes,especially in the civiland political domains 103,145,146,147,148,149.Figure A.8:AI Cyber operationsAI techniques may be used to digest largedatasets of textual data to conduct sentiment andsocial med

298、ia analysis.Unsurprisingly,such largedata collection exercises can be leveraged politi-cally and militarily to influence storytelling,nar-rative creation and present emerging threats to pol-itics 145,146.Some AI techniques,such asbehavioural research advances in natural languageprocessing(NLP)and AI

299、-enabled synthetic text,video and audio(e.g.fake social media posts)inautonomous agents(e.g.bots)use disruptive deep-fake information as a statecraft tool to shape public perception to convenient geopolitical or militarysignificant storytelling.Hence,demonstrating how content-based AI-enabled inform

300、ation is weaponisedand presents a threat 145,146,149.Further applications of embedded AI in autonomous vehicles(AVs)provide the intelligence behindthe decision-making process,planning and execution of tasks 135.For instance,vehicle control andhuman assistance depend on the level of entrusted autonom

301、y.Standard or partial AVs perceive the contextrelying on advanced sensor networks and interact within it,signalling the need for human intervention.Whereas,highly automated vehicles(HAVs)depend on AI to operate complex data-intensive decision-making functions required to understand the surroundings,

302、plan for subsequent actions,and manoeuvre135.For instance,Project Maven embodies and weaponises the convergence of AI algorithms anddrones in support of mission intervention 134.Figure A.9:Embedded-AI in Autonomous VehiclesAIs ability to perceive and interact within anenvironmentcomesfromadvancesinc

303、omputersci-ence,sensors and geospatial technologies,human-logic or interpretations,and AI techniques.Forexample,all improving performance of deep neu-ral networks(DNNs)algorithms used to recogniseand train on selected images,objects detection,faces identification,NLP)/(NLU),translation,andso forth 1

304、35.Drawing upon DNNs advancements,NLP hasreached human-level performance with NLU.Thisevolution contributes to the next step of text to image synthesis,which consists of systems that cangenerate images based on textual data.Generative AI is considered one of the most challenging areasof AI 145.In ad

305、dition,breakthroughs in high-resolution image recognition,spanning from facialrecognition to more subtle facial expression,enable biometric identification through facial features andemotion analysis 145.Furthermore,use cases in the medical domain point out AIs high-resolutionimage recognition capabi

306、lities,which allow it to identify tumours and cancer cells 143,150.Thereon,raising expectations for military or counter-terrorism applications 145.Human-Machine SymbiosisProficient human-AI teaming is a cornerstone for future AI applications.Mastering the harmonisation ofAI with human-driven systems

307、 to create an effective psycho-social-technical collaborative system will bea game changer.34Figure A.10:Human-Machine SymbiosisResearch on human-machine synergies is mov-ing forward rapidly,focused on developing ex-plainable an observable AI,understanding trust-building and trustworthiness,entruste

308、d authority,and ensuring confidence(validation and assurance)in the operationalisation of AI-enabled systems151,152,103.Increasing scientific understanding of the hu-man mind will accelerate human-machine inter-faces as a pervasive disruptor.Complementary,advanced NLP progress on language processing

309、and Learning enables system development to process complex inputs and engage in human-like conversa-tion while developing a collaborative intelligence among cohorts of agents.Still,XAI,trust,and testing,validation,verification&assurance(TVVA)remain critical challenges for AI in human-machinesymbiosi

310、s as for any defence and security environment.Explainability of AI intrinsically relates to trust,which is the Achilles heel of human-machineteaming.Understanding how to generate a trusted relationship between humans and AI and what itmeans for AI to trust human judgement 42,153 represents a difficu

311、lt technical and psychologicalchallenge.Central to building trust is understanding(at some level)and explaining how an algorithmmade a particular assessment and considered it in making this decision.New ML techniques are beingdeveloped to support their assessments and explicitly communicate associat

312、ed constraints and limitations.Other groundbreaking approaches are being explored to improve human interaction with AI.Theobjective is to develop user-friendly interfaces capable of surfacing acquired knowledge on demand.Thesehuman-interpretable ML models are combined with visualisation and communic

313、ation interfaces moreconducive to human understanding,and evaluation 42,106,153,154,155.TVVA also represents a critical aspect of trust,ultimately one of the most challenging aspects of theoperationalisation of AI/ML.R&D is moving forward to create systems engineering methods,standards,assessment to

314、ols and software;to ensure that deployed AI is timely,affordable,trusted and fully missioncapable.This will require real-time tools to guarantee reliability and traceability.In addition,researchinto new continuous learning algorithms requires that agility be balanced with the need for missionaccredi

315、tation 42,153.Research is also focused on manipulating human decision-making by AI 156,157.ResilienceSince AI is being adopted and deployed by both BLUE and RED forces,AI-on-AI engagements arecurrently emerging(mainly in the realm of disinformation).As a result,the resilience of intelligentsystems a

316、nd counter-AI activities become increasingly critical to detect,deflect and limit the impact ofattacks on Alliance AI while undermining adversarial AI-enabled systems,as noted by the NATO ChiefScientists research report(CSRR)on AI1.Adversarial use of AI/ML unveils a collection of tools and technique

317、s designed to defeat or obstructintelligent systems currently impacting military sensors,weapon systems,and decision support systems(e.g.Intel,Socint).The interference with,and jamming of,inputs modifies outputs imperceptibly to thehuman eye,thereby interfering with daily and future operations 158.I

318、n contrast,tightly dependent onthorough intelligence on adversary systems 159,160,161.Potential uses of advanced ML in researchare identified as:(i)countermeasures against detectors in next-generation sensor systems;(ii)obstructionof ML-based remote sensing data analysis.The importance of exploring

319、AI countermeasures is largely shared among the S&T and militarycommunity as it represents a technology prime for increased R&D.As Alliance and competitor forcesprogressively use AI,AI countermeasures become increasingly critical to explore to understand howto protect and exploit AI-based systems 159

320、,160,161.Like any operational domain,there is a need1To access the report visit -Publications,NATO Chief Scientists Reports section35Chapter A.Artificial Intelligenceto detect,deflect and limit the impact of attacks on Alliance AI.AI robustness set of tools developedfor TVVA and those used to create

321、 adversarial AI countermeasures are being extended into a real-timedefence of BLUE-AI systems,including the need to harden AI against deception by adversaries.Figure A.11:Countering AI SystemsCertainly,counter-AI in general,counteringadversary AI and associated decision processesthrough AI manipulat

322、ion and deception is a diffi-cult technical challenge yet to be solved.Similarly,detecting AI manipulation is not straightforward.On the contrary,AI-on-AI conflict is not just a pos-sibility but is a complex,evolving reality in cogni-tive warfare and disinformation 149.Therefore,AI systems engagemen

323、ts will become increasinglycommon,thus requiring thorough identification ofadversarial AI,model AI responses,timely andefficient weaknesses assessments,and counter-AI strategies development 162.Military ImplicationsArtificial Intelligence is currently transforming warfare.According to NSCAI 2021,103

324、:AI applications are transforming existing threats,creating new classes of threats,andfurther emboldening state and non-state adversaries to exploit vulnerabilities in our opensociety.AI systems will extend the range and reach of adversaries.just as the missileage and terrorism bring threats closer

325、to home.Because of AI,adversaries can act withmicro-precision but at a macro-scale and with greater speed.They will use AI to enhancecyber attacks and digital disinformation campaigns and to target individuals in new ways.AIwill also help create precisely engineered biological agents.And adversaries

326、 will manipulatethe AI systems we will rely upon.Blues comparative edge depends on surgical leverage,exploitation and adoption of AI.Besides,theresulting AI capability will underpin the Alliances ability to compete,deter,and win across domains163.BLUEAI will significantly impact Alliance military ca

327、pabilities and forces.This impact will predominatelyoccur through Chinas new concept of intelligentized warfare and the control of the enemys will 164;embedded AI in other associated technologies such as virtual/augmented reality;quantum computing;autonomy;modelling&simulation;space;materials resear

328、ch;manufacturing&logistics;and big dataanalytics 108.In addition,AI will transform nuclear,aerospace,cyber,materials and biotechnology.Some argue thatthese effects will have a strategic impact on the same scale as the introduction of atomic weapons 165.Furthermore,an over-reliance on AI systems will

329、 introduce significant new vulnerabilities and usher inan adversarial AI arms race 103.Figure A.12 offers a summary of emerging AI threats BLUE shouldconsider while adopting,implementing or fielding any AI solution.Even though AI presents risks,potential defence applications should continue consider

330、ing the appro-priate level of human control and accountability.Nevertheless,many of these scientific and technologicaldevelopments will inevitably impact operational,strategic and enterprise capabilities/business functions.EnableEnterprise Management:NATO requires more efficient and effective proces

331、ses for enterpriseresource management(investment and business planning,program performance and risk manage-ment,strategic transformation and improvement initiatives,strategic readiness management,andstrategic management practice).These improved processes will be based on advanced analytics,and36Figu

332、re A.12:Emerging threats in the AI era:How AI is transforming the threat landscape(adapted from103.evidence-based decision-making 166.AI can assist in cost analysis,assessing economic impactsand drivers,and providing timely,evidence-based decision support for finance.Applied AI/Logistics:The support

333、 of AI technologies could increase the speed and responsivenessof logistics planning in future operations.Symbiotic human-machine integration:through meaningful human control represents a mustto have to extract value from technologies such as AI,big data analysis,and the application ofautonomy,which can enable NATO to operate efficiently and respond dynamically to changingcircumstances,supporting