Assistant Professor Ismini Lourentzou has received a National Science Foundation (NSF) CAREER award to develop the next generation of embodied AI agents, systems that can reason, explain, and adapt as they act in the physical world. This prestigious award is given in support of early-career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization. Lourentzou's project, "Shaping Embodied Intelligence Through Language-Guided Introspection," will be supported by a five-year, $600,000 grant from the NSF.
As AI systems move into factories, hospitals, and other real-world physical environments, safety, reliability, and effective human-AI collaboration become increasingly important.
"Today's AI agents can be powerful, but they often fail in ways that are difficult to detect, explain, and correct," said Lourentzou. "They may misinterpret an instruction, fail to recognize when their reasoning is flawed, struggle to recover from errors, or persist in unsafe or ineffective behavior. In the physical world, these failures affect safety, productivity, and trust."
Lourentzou's project aims to build self-regulating AI agents that can recognize uncertainty, explain their decisions using past experience, and improve their understanding of the world through corrective feedback. The project will establish a new foundation for embodied intelligence by using language as a mechanism for self-monitoring, connecting perception, reasoning, memory, and world models as AI agents act in dynamic physical settings.
"Specifically, this project explores how language can serve as an internal self-regulation mechanism, similar to how humans use inner reasoning processes," she said. "The broader goal is to make future physical AI systems more trustworthy and useful in real-world settings."
The educational component of the NSF CAREER award will translate this research into new learning experiences for students, including research opportunities, classroom modules, and community engagement activities designed to broaden AI literacy and strengthen workforce readiness in trustworthy AI.
Lourentzou's research focuses on multimodal machine learning, primarily vision-language models, generative modeling, embodied AI, and grounded reasoning for physical and interactive environments. She leads the Perception and LANguage (PLAN) Lab. Lourentzou earned her PhD in computer science from the University of Illinois Urbana-Champaign.