The increasing impact of Artificial Intelligence (AI) necessitates the development of AI literacy in education. The AIEd community emphasizes the importance of training teachers and students to understand AI's fundamentals, not just trending tools. This involves teaching core concepts like perception, representation, reasoning, learning, and the impact of AI. Current AI literacy initiatives mostly focus on machine learning, particularly supervised learning, while other crucial areas like perception and reinforcement learning are less addressed. This gap highlights the need for effective teaching resources, especially those incorporating real-time interaction and environmental uncertainty. The Robobo Project, utilizing intelligent robotics, is proposed as a solution to this problem. Robots offer a tangible representation of intelligent agents, allowing students to engage with AI concepts in a dynamic and hands-on manner. The project aims to address how to integrate AI education into existing curricula, how educators can effectively teach AI, and what resources can be used at different educational levels. The Robobo Project offers a potential long-term solution to these challenges.

Educational Robotics (ER) has expanded beyond technical universities to encompass general education. Several ER initiatives exist, utilizing robots of varying complexities for different age groups. However, many existing platforms (e.g., Thymio, Fable, LEGO MINDSTORMS) lack features essential for teaching AI, such as advanced sensors, actuators, computational power, and comprehensive AI-focused educational materials. While some platforms address aspects of AI, a comprehensive approach covering all key concepts across educational levels is missing. Existing ER platforms for AI education, like Thymio and Fable, offer some relevant features but lack others, such as a complete set of educational materials aligned with AI literacy recommendations and a robust simulation environment. This gap motivated the creation of the Robobo Project.

The Robobo Project employs a multi-faceted approach combining engineering and educational design research. The Robobo robot integrates a mobile robotic base with a smartphone, leveraging the smartphone's advanced sensors and computational capabilities. The hardware comprises sensors (infrared, encoders, camera, microphone, accelerometer, gyroscope, etc.) and actuators (motors, speakers, LEDs, screen). Software development followed a Unified Process, iteratively adding functionalities. The core software runs on the smartphone, enabling modularity and easy adaptation to different learning objectives. Educational materials include teaching units (TUs) and documentation. The TUs, aligned with AI literacy principles, present challenges for students to solve using the robot, based on Project-Based Learning (PBL). The RoboboSim, a Unity-based simulator, allows students to work in a virtual environment before transitioning to the real robot. Both Scratch (block-based) and Python programming are supported, catering to different educational levels. The Robobo software includes libraries and programming frameworks for various functionalities, including vision, sound, and tactile sensing, and actuation.

The Robobo Project has been validated through extensive use in classrooms across diverse educational levels (secondary school, university, and master's programs) in various European countries. Questionnaires were administered to assess the project's effectiveness. Results show Robobo is perceived as a suitable long-term tool for AI education, effective across different age groups and skill levels. The use of a smartphone as a central component received positive feedback, though some preferred dedicated devices for technical reliability and consistency. The combination of real-world and simulated environments was well-received, allowing students to focus on programming concepts without being bogged down by hardware issues initially, then applying those skills to the real world. At the secondary school level, teachers rated machine learning and social impact of AI as highly relevant topics, highlighting the effectiveness of the project's design to teach core concepts even without advanced AI expertise. The computer vision capabilities and human-robot interaction (HRI) aspects of Robobo were highly valued by students. At the university level (degree and master's programs), the project's ability to engage students in practical problem-solving through realistic challenges was positively received. The use of Python and the accessibility of the documentation were also key factors. Advanced students successfully integrated advanced algorithms and libraries, demonstrating the system's adaptability. Master's students found the Robobo platform suitable for exploring advanced AI concepts, including computer vision, path planning and control architectures, with the simulator providing a realistic, flexible and effective environment to learn.

The findings demonstrate the Robobo Project's success in addressing the research questions. The project provides a robust and adaptable tool for long-term AI education, bridging the gap between theoretical concepts and practical application. The positive feedback from both students and teachers underscores the effectiveness of the project's methodology and materials. The successful integration of a smartphone-based platform demonstrates cost-effectiveness and accessibility, while also ensuring the project's adaptability to future technological advancements. The comparison with Thymio and Fable highlights Robobo's unique strengths in providing a comprehensive AI education experience. Robobo’s success lies not just in hardware but in its comprehensive software ecosystem, aligned with AI literacy recommendations, and its adaptable teaching materials.

The Robobo Project offers a viable and effective long-term solution for integrating AI literacy into education. The use of intelligent robotics, coupled with carefully designed software and educational resources, provides a compelling learning experience across different educational levels. Future work could explore expanding the Robobo's capabilities, developing new teaching units, and further refining the simulator to enhance the learning experience. The findings support the wider adoption of robust educational robotics platforms like Robobo to equip future generations with crucial AI literacy skills.

The study's scope is limited by the specific geographic regions and educational institutions involved. The sample sizes, while significant, might not fully represent the broader population of students and educators. Further research could involve larger-scale studies across diverse contexts to enhance generalizability. While the Robobo system is designed for adaptability and long-term use, the reliance on smartphone technology might introduce certain constraints, such as variability in smartphone capabilities and the need for updates to maintain compatibility with evolving technology.