生成式人工智能军事应用中的人为因素与人类系统整合

Frank O. Flemisch With support from: Michael Preutenborbeck, Laurenz Burlage, Alexander Ripkens Frank.flemisch@fkie.fraunhofer.de Frank Flemisch 1967 – ca. 2044 Human Factors and Human Systems Integrationof Military Applications of Generative AI Frank O. Flemisch With support from: Michael Preutenborbeck, Laurenz Burlage, Alexander RipkensFrank.flemisch@fkie.fraunhofer.de Natural Intelligence & Human Systems Integration Schöningen 1995 Engineering: TechnologyHuman Factors: Human + TechnologyHuman Systems Integration: Techno + Human + Orga + Env. Human & Artefacts: Human – Nature Cooperation Human-“AI“-Past: Talos (750 – 650 BC) "a brassrobotmeasuring 30m high, who was forged by the god Hephaestus" Human-"AI" past: Oracle of Delphi (800 BC – 400 AD) -Complex questions by priest Phytia (for the rich)-Simple questions "Yes-No" by sub-priest (for the poor)-medèn ágan,gnôthiseautón Human Intelligence: Dialectics Georg WilhelmFriedrichHegel(1770 – 1831) 1983: Near extinction of homo sapiens, saved byStanislav Petrov History of AI History of AI History of AI Menetekel "Homo Deus" Is humanity soon at the end … … or is a new human narrativeneeded? In "Homo Deus", Israelihistorian Yuval Noah Harari sketches ableak vision of the technological age. By Adrian Lobe If AI is like Fire … Wendell Wallach, Keynote at ReAIM Conference 02/2023 If AI is like Fire … History of AI HSI Past: Dialectics The gap between human and technology Tension Field: Human – Technology / Bio & Technosphere Tension Field: Intelligent – Non-intelligent Non-intelligent Tension Field: Want – Can – Should - Do Tension Field: Fast – Slow Tension Field: Critical – Optimistic Media attention Tension Field: Human – Technology Tension Field: Intelligent – Non-intelligent Cooperative Intelligence: Example H-Metaphor Horseman and Horse: H(orse)-MetaphorTight Rein Cooperative Automation & Cooperative AI:Shared + Traded Control = Cooperative Control Levels of Automation & Levels of AI? Human-AI-Teaming / Hybrid Cooperate Intelligence: Joint Cognition! Optimizing the workload of agents in theidentification of air targets −Improvement of situational awareness−Reduction of errors−Reduction of time needed for management −Supervision of the interaction between theoperator and the Human-Machine-Interface (HMI)−Dynamic adjustment of the HMI to allowfor appropriate (time & situation) executionof the actions Do we have the right metaphor(s)? Anthropomorphic Metaphorfor AI & Automation Zoomorphic Metaphorfor AI & Automation Anthropo-Zoomorphic Metaphorfor AI & Automation Centaur(e.g. Centaur Chess,in Human-machine Teaming Human Machine Symbiosis (Licklider 1960, Gill1996)Cooperative Systems (Hoc, Pacaux-Lemoine1997 ff) „Co-Pilot“, „Electronic Assistant“ etc.Onken et al. 1987ffAmalberti et al. 1989 ffFlemisch et al. 1997 ff If AI is like Fire…we will need Fire protection + Firefighters Wendell Wallach, Keynote at ReAIM Conference 02/2023 Intelligence? Cognition! Cybernetics: Systemic thinking about Joint Cognitionand Acting Balanced Thinking about Human-Machine Systems:Dilemma Model Human Use Space: Dilemma Model Holistic Thinking about Human-Machine SystemsHuman Holistic Bow-Tie Model of Human-Machine SystemsHuman Holistic Bow-Tie Model of Human-Machine SystemsHuman Holistic Bow-Tie Model of Human-Machine Systems Human Holistic Bow-Tie Model of Human-Machine Systems Holistic Bow-Tie Model of Human-Machine Systems Tension Field: Fast – Slow System 2: Conscious, slower Tension Field: Fast – Slow From Levels to Layers of Human-AI Cooperation From Shared to Cooperative Control of AI “One fits all?”: Common (Ground) Control Station Transparency and Meaningful Human Control are key! Flemisch, F.;Baltzer, M.;Abbink, D.; Siebert, L.; Diggelen, J.; Draper, M.; Boardman, M.; Pacaux-Lemoine, M.: Towards a Dynamic Balance between Humans and AI-basedSystems: System of system Perspective on Ability, Responsibility, Authority, Autonomy, Meaningful and Effective Control, andAccountability;Handbook on Meaningful HumanControl;Ed. Siebert, L.;Abbink, D.; Edward Elgar Research Handbook; 2022/23 Human-AI Integration for conflicts Tension Field: Critical – Optimistic Tension Field: Critical – Optimistic Bridging methods for a balance betweenhumans, technology/AI, organizations and environment If AI is like fire, water, horses: Let's tame it together! Without-Getting kicked-Getting burned-Drowning We need to speed up, evil is not waiting: e.g. Putin Human Factors and Human Systems Integrationof Military Applications of Generative AI Frank O. Flemisch With support from: Michael Preutenborbeck, Laurenz Burlage, Alexander Ripkens and Ellen Klinkel Frank.flemisch@fkie.fraunhofer.de