Participating in the design of AI robot assistants through prompt engineering

As social robots with embodied intelligence driven by multimodal large language models, or AI robot assistants, become increasingly accessible, educators require guidance on utilizing these emerging technologies in ways aligned with their pedagogical vision and ethical principles. This conceptual paper explores how prompt engineering can enable educators to intentionally participate in the design of AI robot assistants in learning contexts. Drawing on the conceptual frameworks of zones of possibilities, embodied learning approaches, and social presence, we argue that prompts can be systematically developed to shape how AI robot assistants present content, allocate attention, generate physical movements, and adapt to learners. We illustrate how the IDEA Framework for Prompt Engineering can guide stakeholders in crafting prompts that reflect sound pedagogical principles while prioritizing ethical considerations. Rather than adopting a dichotomous perspective on AI robot assistants, this paper positions prompt engineering as a participatory practice through which educators can design AI behaviors to empower learners.