Robertiño: A Low-Cost Robot for Executing Large Language Models and Convolutional Networks to Enhance Early Education in Underserved Communities

This paper presents Robertiño, a low-cost robotic assistant designed to enhance early childhood education through the use of accessible artificial intelligence technologies. The system integrates a mobile application, an open-source server architecture, and a set of convolutional neural networks and natural language processing models. Robertiño interacts with children through multisensory experiences that include image and voice recognition, real-time feedback, and physical gestures. The robot is designed using recycled and low-cost hardware components, making it suitable for under-resourced educational environments. A pilot study involving 55 children and 3 expert teachers was conducted to evaluate the system. While the internal consistency of the children’s survey was limited due to developmental factors, the expert evaluation revealed substantial consensus regarding Robertiño’s pedagogical, technological, and inclusive value. These findings support the viability of deploying low-cost, AI-driven educational tools in early learning settings.