BEYOND SAFETY FORUM

Gianpiero Negri is Fellow for Robotics Functional Safety at NXP Semiconductors, where he drives design compliance strategy, AI and functional safety frameworks, and global safety governance for large-scale robotic systems. Prior to joining NXP, he served as Global Head of the Robotics Safety Center of Excellence at Amazon Robotics, and before that as Principal Manager in EU R&D Robotics Safety Engineering at Amazon Robotics Advanced Technology, EU Advanced Technology Machine Safety Manager, and Global Functional Safety Manager at CNH. He holds a PhD in Applied Mathematics (Robotics AI Safety), a Post-MSc in Industrial Robotics and Functional Safety (CNR-ITIA), and an MSc in Computer Engineering (summa cum laude, Università di Salerno), and is certified in functional safety (CMSE) and AI/ML (IBM and Stanford).

Gianpiero is an active contributor to international standardization, serving on ISO TC299, ISO/IEC JTC1 SC42, and CEN/CENELEC JTC21 committees. He has presented at leading safety and robotics conferences including the MIT STAMP/STPA Workshop (Cambridge, MA), SAFECOMP/WAISE, ICRA, IRSC, NVIDIA GTC, etc. Real-Time Safety Engineering for Advanced Robotics at Scale

Real-Time Safety Engineering for Advanced Robotics at Scale

As robotic systems become more adaptive and complex, safety can no longer rely on one-time validation before deployment. This talk explores how real-time monitoring, adaptive control, and online verification are needed to continuously ensure safety during operation. It also highlights how established safety principles are challenged when scaling from single certified systems to large fleets deployed across multiple sites. Maintaining consistency, robustness, and trust in safety at scale becomes a key issue for next-generation robotics.

Dr. Sebastian Hallensleben is the Chair of CEN-CENELEC JTC 21 where European AI standards to underpin EU regulation are being developed, and co-chairs the AI risk and accountability work at OECD. As Chief Trust Officer at Resaro, he works towards drilling down to ground truths about capabilities of AI systems and towards standards for measuring use-case-specific AI quality characteristics, including in ETSI MTS. Sebastian is the initiator and Programme Chair of the Digital Trust Convention and is Principal Advisor Digital Trust at KI Park. - Previously, Sebastian headed Digitalisation and Artificial Intelligence at VDE Association for Electrical, Electronic and Information Technologies. He focuses in particular on operationalising AI ethics, on characterizing AI quality and on building privacy-preserving trust infrastructures for a more resilient digital space.

Describing and measuring AI quality: the new frontier

The AI scene is brimming with bold claims about AI as a technology but scarce on hard measurable facts about the characteristics of AI systems - both with regards to the handling of safety, security and fundamental rights risk, and with regards to performance in delivering useful outcomes. In this talk, Sebastian Hallensleben presents how this is gradually changing through standardisation, drawing on current work in CEN, CENELEC and ETSI.

Prof. Dr. Nico Hochgeschwender - is Professor of Software Engineering for Cognitive Robots and Systems at the University of Bremen, where he leads the SECORO research group. His research lies at the intersection of AI-enabled robotics and software engineering, with a focus on safety and dependability, validation and evaluation, benchmarking, and domain-specific modelling for robotics. He coordinates the MSCA Doctoral Network CAVECORE on continuous automated validation and evaluation of cognitive robots and is Co-Speaker of the Robotics Institute Germany (RIG) Technical Cluster on Safety for AI-enabled Robotics. From 2019 to 2025, he served as Co-Chair of the IEEE RAS Technical Committee on Software Engineering for Robotics and Automation.

From Test Cases to Test Spaces -- Scenario-Based Validation for Safe Autonomous Robots

Autonomous robots operate in increasingly complex and variable environments, making it almost impossible to anticipate all relevant situations during the validation process. Instead of relying on individual test cases, this talk explores how explicit scenario models can define structured test spaces from which concrete and reproducible robot experiments can be systematically generated and executed in both simulations and real-world settings. I will discuss how such executable test spaces can help relate operating conditions and safety requirements to observed robot behavior and what challenges remain regarding coverage and the use of test results as evidence for safety assurance.

...