Scope and Aims. The increasing deployment of intelligent systems (IS) [1] across critical domains necessitates a deep understanding of their interaction with social, ethical, and legal frameworks. These systems are inherently socio-technical, requiring an interdisciplinary approach that integrates technical design with social and human sciences, domain expertise, and industry practice. This meeting is designed as an interdisciplinary seminar to facilitate this convergence. The objective of this Shonan meeting is to establish a foundational research roadmap for developing trustworthy Socio-Technical Intelligent Systems. We aim to move beyond abstract discussions by grounding our work in concrete industry use cases. This meeting will gather researchers, social and human scientists, domain experts, and industry professionals to identify challenges, define requirements from normative principles, and brainstorm methodology to design systems that are both effective and socially responsible. To ensure a productive and focused discussion, participants will engage in pre-meeting preparation based on their expertise: (i) Industry Participants: Will propose and detail a concrete industry use case; (ii) Social and Human Scientists: Will identify relevant norms and legislation that apply to the proposed use case; (iii) Domain Experts (e.g., Doctors for a medical use case): Will identify relevant domain-specific norms and professional standards; and (iv) Computer science researchers: Will review existing system requirements, explore analysis and modeling techniques (e.g., digital twins and related decision making systems), define challenges, and synthesize the normative principles into formal, actionable requirements.

Topics. The meeting expects the participants to contribute to one or more of the following research topics: What are the core components and architectural patterns of Socio-Technical Intelligent Systems? How can concrete industry use cases be effectively modeled and analyzed to surface socio-technical challenges? How can relevant social norms, legislation, and ethical principles be systematically identified and translated into formal system requirements? What methodologies can be used to integrate domain-specific norms (e.g., medical, legal) into the design and verification of intelligent systems? How can techniques like Digital Twins or formal analysis be leveraged to analyze existing system requirements against established normative principles? What are the major challenges and opportunities in defining and enforcing normative requirements for intelligent systems (e.g., bias, fairness, transparency, accountability)? How can the definition of system requirements be informed by the interdisciplinary perspectives of social science, domain expertise, and engineering? What is the research roadmap for moving from normative principles to verifiable and trustworthy Socio-Technical Intelligent Systems?

Format. The meeting will be organized to maximize interdisciplinary integration and ensure the production of concrete research outputs. The structure, while maintaining some flexibility, will follow these key stages: Keynotes and Mini-Tutorials (Days 1–4): Foundational talks delivered throughout the first four days to establish a shared understanding of socio-technical systems, normative frameworks, and foundation-model challenges. Synthesis of Pre-Meeting Outcomes: Participants will briefly present their pre-meeting preparation (industry use cases, relevant norms and legislation, domain-specific standards, or technical analyses) along with what they expect and need from this multidisciplinary event. This synthesis will serve as the basis for identifying common themes, gaps, and priority challenges. Guided group discussions. We will seek to guide the formation and discussions into groups in three phases. 1) Idea Speed Dating in which we get to meet and find wild possibilities, focusing on what each participant can provide (e.g. tools or framework) and the challenges they need help with. 2)Deepening the Relationship in which we try to find more enduring matches and generate many possible ideas and sharing them with the group. 3) Getting Serious in which we start to refine the ideas and put together our plans for the future. Group Discussions and Reporting:Participants will be organized into focused, interdisciplinary working groups aligned with the key challenges and research opportunities identified during the synthesis session. Groups will report their findings in joint sessions to consolidate knowledge, reconcile perspectives, and establish shared problem definitions. NIER Paper Drafting (Final Day): The meeting will culminate in a collaborative writing session focused on producing a draft NIER (New Ideas and Emerging Results) paper, capturing the research roadmap, methodological propositions, and open challenges identified during the seminar.

Relations to Previous Shonan Seminars. This proposed Shonan seminar builds on earlier meetings such as Engineering Trustworthy Foundation Models (No. 227), Science and Practice of Engineering Trustworthy Cyber-Physical Systems (No. 178), and Formal Methods for Trustworthy AI-based Autonomous Systems (No. 055), which primarily addressed technical foundations, verification, and engineering challenges of trustworthy AI and autonomous systems. In contrast, this seminar narrows and deepens the scope by focusing explicitly on Socio-Technical Intelligent Systems interacting with humans in society, where such interactions can lead to social, legal, ethical, and empathetic harms. Ensuring trustworthiness in this context requires that normative principles be systematically integrated into system engineering, and translated into explicit requirements, architectures, and assurance processes, grounded in concrete industry use cases. The seminar is also complementary to Social Explainable AI (No. 200): while sharing an interdisciplinary perspective, it shifts the focus from explanation design to the engineering of norm-aware and accountable intelligent systems from the outset, with strong involvement from industry and domain experts.

[1] Anetta Jedličková. Ethical approaches in designing autonomous and intelligent systems: a comprehensive survey towards responsible development. AI & SOCIETY, 2024.