Expert Insights: Advancing AI Systems in Engineering

AI and R&D Karen Grothe karen@karengrothe.com www.incose.org/IW2021

Artificial Intelligence (AI) Systems WG Purpose: Identify and communicate emerging AI technologies that appertain to industries of the Future. The AI Systems WG seeks to establish INCOSE as a recognized resource of expertise for research, development, and application of AI systems. 1 General Goals: Identify the needs of the international AI community (industry, academia, government) which are well-suited for contributions by INCOSE, and to provide expertise across SE functions and lifecycle management that can be used by industry to promote the development of AI Systems. 2 Scope: This WG addresses the use of AI within Systems and using AI as part of the Systems development. The WG will encourage research to be conducted by members or via collaboration with other organizations. 3 20-mars-25 www.incose.org/IW2021 2

Specific WG Goals and Objectives Specific Goals 1) Identify and communicate emerging AI technologies that can be applied to the engineering of systems (AI for SE) 2) Develop and communicate advances in SE methods needed to effectively engineer systems with embedded AI (SE for AI) Research Objectives Explore Human-AI collaboration Evaluate Safety and Security of AI systems Measure and evaluate AI technologies through standards and benchmarks Understand and promote workforce development and STEM initiatives Contribute to public-private partnerships and affiliations to accelerate advances Establish best practices for using AI techniques in Systems and Systems Engineering 20-mars-25 www.incose.org/IW2021 3

AI-SE Primer Started at IW2020 under the Future of Systems Engineering (FuSE) Initiatives 4 Sections: What is AI and Explainable AI? AI for Engineered Systems Engineering AI with Systems Engineering AI and the Future of Systems Engineering Initial Draft Target: IS2021 20-mars-25 www.incose.org/IW2021 4

AI Ethics: New and Old Joint discussion held by Social Systems, Complex Systems, and AI WGs Why is it such a hot topic? (What makes it special?) Fear of the unknown and unleashing the devil Power and control Spread of misinformation Unique Ethics of AI Why is it important to be grounded in ethical concepts? What it is and what it isn t Business and behavior. From an engineering ethics perspective, technology and engineering artifacts are ethically neutral. Engineering ethics Decision making of the engineer. Influences that may take place without the engineer being conscious of it. 20-mars-25 www.incose.org/IW2021 5

AI Ethics Selected Chat Quotes Humans tell the system what is right and wrong thus humans are responsible for the ethics. There s a difference between bias in data and bias in algorithm. Bias in AI can be in the data that the system is trained with, or in the world itself. 20-mars-25 www.incose.org/IW2021 6

Systems Engineering in Early Stage Research & Development WG (ESR&D WG) Purpose: Provide an open forum for the development, application, and dissemination of systems engineering principles, best practices, and solutions to scaling systems engineering applications to ESR&D projects so that the level of systems engineering effort is commensurate with risk and that the anticipated ESR&D outcomes are achieved. 1 Goals: To provide knowledge, guidelines, and frameworks for the application of systems engineering in ESR&D 2 Scope: Focus on activities at Technology Readiness Levels (TRLs) 1 5 3 20-mars-25 www.incose.org/IW2021 7

ESR&D WG Current Activities WG chartered in 2020 Desire a framework that contains guidelines and processes for the right and right-sized tailored SE Research and ESR&D is based in principles of expansionism - continuous flow of ideas and experiment. Transition requires reductionism breaking down a system into the sum of its parts Framework must accommodate both expansionism and reductionism 20-mars-25 www.incose.org/IW2021 8

Example Frameworks Sandia National Laboratory Based on Risk, Industry, Standards, and Project Type Identified core set of practices with differing levels of rigor based on intrinsic risk Los Alamos National Laboratory Based on Risk, Standards, and Project Type; called the approach Mission Assurance Identified core set of practices with differing levels of rigor based on intrinsic risk Pacific Northwest National Laboratory Based on Risk, Industry, Standards, and Project Type Define SE triggers which drive implementation of SE (informal, semi-formal, formal) 20-mars-25 www.incose.org/IW2021 9

Other ESR&D WG Activities Publications Implementing Systems Engineering in Early Stage Research and Development (ESR&D) Engineering Projects IS20 30Th Annual INCOSE International Symposium July 18-23, 2020 Perceived Conflicts of Systems Engineering in Early Stage Research and Development Draft for INCOSE InSight August 2021 IS21 July 17-22, 2021 Panel Submission and Preparation 20-mars-25 www.incose.org/IW2021 10