Reevaluating the Trustworthiness of Autonomous Systems
Hong Kong, July 18, 2025 / AP News / -- As autonomous systems accelerate their deployment in industries such as transportation, logistics, and healthcare, the goal of "zero accidents" and "zero downtime" should not only be an ideal target but also a fundamental premise. Under these circumstances, the trustworthiness of end-users remains a crucial factor in industry development.
The "trust gap" facing autonomous systems
Autonomous systems have brought unprecedented efficiency, but their application core relies on trust. This trust originates from the system's ability to run safely, reliably, and stably in dynamic and varied environments. Traditional safety systems focused on compliance are no longer sufficient to address complex and dynamic actual deployment environments; we need to achieve a paradigm shift from "compliance-based security" to "trustworthy adaptive security."
TÜV Süd Group's (hereinafter referred to as "TÜV Süd") patented Adaptive Safety & Security System (AS3) is designed to bridge this "trust gap." The AS3 system integrates two major core technologies: on the one hand, it combines TÜV Süd's professional expertise in adaptive safety systems; on the other hand, it integrates Hong Kong Applied Science and Technology Research Institute's (ASTRI) cutting-edge technology based on model-based systems engineering (MBSE) for digital-physical twin design and development. Through real-time environmental sensing intelligence, autonomous systems can predict, adapt, and run reliably in high-risk scenarios. This collaboration emphasizes the parties' joint commitment to innovative technologies such as autonomous driving and robotics -- in the engineering support system, safety and trust will always occupy a core position.
Key technology highlights
In-real-time network physical risk perception: proactively identifying potential threats from both physical and digital worlds.
Dynamically situational awareness reasoning: breaking through preset rule limitations to flexibly analyze complex environments.
Multidimensional intelligent learning mechanism: actively learning in actual machine-machine, human-machine, and machine-environment interactions rather than relying solely on static data.
Implementable operational insights: real-time detection of system anomalies, performance degradation, and seizing opportunities for optimization before system downgrade.
Transparent decision-making mechanism: enhancing accountability and trust-building by providing explainable artificial intelligence outputs.
TÜV Süd and the Hong Kong Applied Science and Technology Research Institute will soon jointly host a thematic event -- "Trustworthiness and Sustainability: The Two Pillars of New Technologies" -- under the premise that safety, reliability, and environmental management are intertwined. We invite industry elites to explore how to build a feasible framework for deploying autonomous systems together. Further details will be announced separately.