Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Navigating the sociotechnical labyrinth: Dynamic certification for responsible embodied AI

Georgios Bakirtzis, Andrea Aler Tubella, Andreas Theodorou, David Danks, Ufuk Topcu

TL;DR

The paper addresses governance challenges posed by embodied AI, whose dynamic, context-sensitive behavior outpaces static regulatory frameworks. It introduces dynamic certification as an adaptive, iterative regulatory approach that continuously tests and expands an AI system's permissible capabilities using living models, formal methods, and real-world monitoring. The framework is layered to incorporate sociotechnical context, use-case-specific requirements, and system-level testing, with a concrete action path demonstrated through an urban delivery robot scenario. While offering a promising route to safe, ethical deployment, the authors highlight substantial open problems—technical, sociotechnical, and policy-related—that require multidisciplinary research before dynamic certification can be broadly adopted.

Abstract

Sociotechnical requirements shape the governance of artificially intelligent (AI) systems. In an era where embodied AI technologies are rapidly reshaping various facets of contemporary society, their inherent dynamic adaptability presents a unique blend of opportunities and challenges. Traditional regulatory mechanisms, often designed for static -- or slower-paced -- technologies, find themselves at a crossroads when faced with the fluid and evolving nature of AI systems. Moreover, typical problems in AI, for example, the frequent opacity and unpredictability of the behaviour of the systems, add additional sociotechnical challenges. To address these interconnected issues, we introduce the concept of dynamic certification, an adaptive regulatory framework specifically crafted to keep pace with the continuous evolution of AI systems. The complexity of these challenges requires common progress in multiple domains: technical, socio-governmental, and regulatory. Our proposed transdisciplinary approach is designed to ensure the safe, ethical, and practical deployment of AI systems, aligning them bidirectionally with the real-world contexts in which they operate. By doing so, we aim to bridge the gap between rapid technological advancement and effective regulatory oversight, ensuring that AI systems not only achieve their intended goals but also adhere to ethical standards and societal values.

Navigating the sociotechnical labyrinth: Dynamic certification for responsible embodied AI

TL;DR

The paper addresses governance challenges posed by embodied AI, whose dynamic, context-sensitive behavior outpaces static regulatory frameworks. It introduces dynamic certification as an adaptive, iterative regulatory approach that continuously tests and expands an AI system's permissible capabilities using living models, formal methods, and real-world monitoring. The framework is layered to incorporate sociotechnical context, use-case-specific requirements, and system-level testing, with a concrete action path demonstrated through an urban delivery robot scenario. While offering a promising route to safe, ethical deployment, the authors highlight substantial open problems—technical, sociotechnical, and policy-related—that require multidisciplinary research before dynamic certification can be broadly adopted.

Abstract

Sociotechnical requirements shape the governance of artificially intelligent (AI) systems. In an era where embodied AI technologies are rapidly reshaping various facets of contemporary society, their inherent dynamic adaptability presents a unique blend of opportunities and challenges. Traditional regulatory mechanisms, often designed for static -- or slower-paced -- technologies, find themselves at a crossroads when faced with the fluid and evolving nature of AI systems. Moreover, typical problems in AI, for example, the frequent opacity and unpredictability of the behaviour of the systems, add additional sociotechnical challenges. To address these interconnected issues, we introduce the concept of dynamic certification, an adaptive regulatory framework specifically crafted to keep pace with the continuous evolution of AI systems. The complexity of these challenges requires common progress in multiple domains: technical, socio-governmental, and regulatory. Our proposed transdisciplinary approach is designed to ensure the safe, ethical, and practical deployment of AI systems, aligning them bidirectionally with the real-world contexts in which they operate. By doing so, we aim to bridge the gap between rapid technological advancement and effective regulatory oversight, ensuring that AI systems not only achieve their intended goals but also adhere to ethical standards and societal values.
Paper Structure (6 sections, 2 figures)

This paper contains 6 sections, 2 figures.

Table of Contents

  1. Introduction
  2. AI governance challenges: Towards dynamic frameworks
  3. Dynamic certification
  4. Dynamic certification in action
  5. Open problems in dynamic certification for embodied AI
  6. Conclusion

Figures (2)

  • Figure 1: The dynamic certification for sociotechnical AI systems contains bidirectional, interdependent and layered challenges.
  • Figure 2: Dynamic certification explicitly allows for ambiguities in the initial specifications, uncertainties, and decisions yet to be made. In addition, dynamic certification keeps a decision trail by intertwining modeling and testing throughout the lifecycle. Models are not only represented by code, but also living documents continually updated in response to changing data and play a useful role in recording changing assumptions and specifications. Models, specifications, and tests are continually refined as we better understand real-world contexts.