#1307: Lowering Barriers for Designing AI-Embodied Robots for Learning

This paper presents RoboHearts AI, a web-based platform that enables educators, students, and researchers to easily design and deploy large-language-model (LLM)–powered characters on embodied robots such as Pepper, NAO, and LEGO-based extensions. By combining LLM-driven conversational capabilities with robots’ multimodal features, including gesture, animated movement, gaze, and visual display, RoboHearts AI supports naturalistic, adaptive, and contextualized interactions that extend beyond traditional screen-based chatbots. The paper introduces practical applications across educational and non-educational settings, and proposes two pedagogical scenarios: learning with AI, where robots act as tutors or learning companions, and learning about AI, where students gain AI literacy and deepen their understanding of AI by designing and iteratively refining robot characters. This hands-on demonstration guides participants in character creation and multimodal behavior design. This work lowers barriers to embodied AI adoption and highlights new opportunities for research, instructional innovation, and human-centered learning with AI-embodied robots.