May 8, 2026
ACEWORKS Presents Technology Vision for the ‘Safety-First’ AI Era
ACEWORKS, a specialist in future mobility control and verification, held a company-wide all-hands meeting at its headquarters in Gangnam-gu, Seoul on the 6th, where it laid out its technology vision for the AI era — with safety at the center of the conversation.
The presenter was Jong-geun Lim, Platform Team Leader of the ES (Embedded Systems) Division, who took the audience through a new safety paradigm under the theme "Redefining Automotive HILS in the AI Era."
Lim made the case that the AI "brain" — which operates probabilistically and is therefore inherently fallible — must be kept architecturally separate from the safety monitor responsible for physical control. The logic, he argued, is counterintuitive but important: the more capable AI becomes, the more rigorous the verification of the underlying controllers that support it needs to be. As a path forward, he proposed evolving beyond conventional HILS-based verification toward an AI-driven HILS approach — one that stress-tests systems by injecting tens of thousands of unexpected edge-case scenarios to expose their limits before they are ever encountered in the real world.
The presentation also touched on the growing importance of new safety analysis frameworks as autonomous driving technology matures. Lim argued that conventional component-level failure analysis is no longer sufficient, and that the industry needs to adopt STPA (Systems-Theoretic Process Analysis) — a methodology capable of examining the complex interactions between systems, not just the failure modes of individual parts. The broader aim is to move away from one-time certification toward a model of continuous safety assurance suited to the OTA (Over-the-Air) update era, enabling residual risks introduced by software changes to be monitored on an ongoing basis.
