Recent advances in Large Language Models (LLMs), Visual Language Models (VLMs), and other foundation models present new opportunities for robotics. The 3rd iteration of this workshop focuses on exploring the intersection of these models and robotic systems, highlighting how progress in the AI and computer vision communities can inform and accelerate robotics research. Integrating LLMs and VLMs into robotic pipelines could enable systems that are more explainable, instructable, and capable of generalizing across tasks. However, achieving seamless integration remains a significant challenge, as existing models often lack the grounded understanding required for real-world robotic applications, including knowledge of physical properties, spatial relationships, and temporal dynamics. Closer integration with robotic platforms may help address these limitations, as real-world interaction provides rich sensory observations and physical feedback that can support the development of more robust and physically grounded intelligent systems.

This workshop seeks to establish an inclusive and collaborative forum for professionals, researchers, and enthusiasts to exchange ideas, share experiences, and build connections within the AI and robotics community, with particular emphasis on supporting and connecting early-career researchers. The program will include invited talks, paper presentations, open panel discussions, networking opportunities, and presentations of exclusive results and demonstrations from the CMU Vision-Language-Autonomy Challenge. Five invited speakers will present their research, perspectives, and future directions on topics at the intersection of AI and autonomous systems, spanning areas such as datasets and benchmarks, software infrastructures, visual-language navigation, situated reasoning, robotics foundation models, and related themes.

See AI Meets Autonomy 2025, AI Meets Autonomy 2024 for the previous iterations of our workshop at IROS 2025 (Hangzhou) and IROS 2024 (Abu Dhabi).