Game-Theoretic Guarantees for Human-Robot Systems

Despite their growing sophistication, autonomous systems still struggle to operate safely in uncertain, open-world situations—as highlighted by public skepticism toward early automated driving technologies. Meanwhile, the excitement around generative AI has been tempered by concerns about individual and societal risks from poorly understood human-AI interactions, where existing guardrails often obscure rather than remove underlying pitfalls. Comprehensive safety assurances remain elusive in both domains—but could insights from one catalyze breakthroughs in the other? This talk will demonstrate how bridging AI’s learned representations and control theory’s safety principles lays a strong common foundation for certifiable intelligent systems. After demonstrating how game-theoretic AI unlocks safety guarantees for complex robotics problems, from quadruped robots to urban driving, we will present early evidence that generative AI systems can, in turn, leverage feedback loop analysis to anticipate the future consequences of their actions on human users, with strong implications on alignment and safety. The talk will end with a vision for general human-AI safety filters that monitor interactions and proactively steer them toward safe and beneficial outcomes.