Learning of Robot Safety Policies via Adversarial Synthetic Scenarios
For developers of Physical AI systems, this work proposes a method to discover high-risk edge cases, but it is currently only a conceptual framework without validation.
This paper proposes an adversarial gamification framework for learning robot safety policies, where a Red Team generates hazardous scenarios and a Blue Team refines safety policies to prevent them. The work is ongoing and presents only a problem formulation and solution architecture, with no experimental results.
In this work, we propose an agentic gamification framework for hazard-informed learning of robot safety policies through synthetic scenarios. We model scenario generation as an adversarial game between two agents: a Red Team that explores the space of potential failures by constructing hazardous situations, and a Blue Team that incrementally refines safety policies to prevent them. This iterative process enables efficient discovery of high-risk edge cases that are unlikely to be captured through random simulation or manual enumeration. By combining classical risk modeling with adversarial scenario generation and modern learning paradigms, this work provides a scalable pathway for embedding safety into Physical AI systems operating in complex real-world environments. The paper describes ongoing work. The contribution is a problem formulation and a proposed solution architecture.