Conference Program

This abstract presents a workshop developed with an early childhood education class (children aged 4 and 5) at a private school in Porto Alegre, Rio Grande do Sul, Brazil. The pedagogical proposal integrates research, maker culture, educational robotics, and Artificial Intelligence (AI), drawing on an approach inspired by Creative Learning (Resnick, 2017).

The project emerged from the children's investigative question, “What is automatic?” Initially, the group developed hypotheses about what might be automatic in the human body, identifying involuntary actions such as blinking, heartbeats, and hiccups (Rinaldi, 2024). The investigation then shifted to the built environment, culminating in the question: “Is the elevator automatic?” From this problem, the children collaboratively designed and constructed a manual elevator, exploring its mechanisms and modes of operation. The desire to transform it into an automatic one inaugurated a new phase of the process: the introduction of Artificial Intelligence (AI) (Russel & Norvig, 2010) as a concrete possibility for experimentation.

We then conducted a workshop to discuss, in accessible language, introductory notions of automation and basic ideas related to AI through hands-on, playful, and embodied activities. Children designed and built personalized toy automata made from laser-cut MDF (Martinez & Stager, 2013), incorporating their own paper-based characters into the mechanisms (Wilkilson et al., 2014). These creative constructions served as the physical interface for subsequent interactions and generated high levels of engagement, requiring careful pedagogical mediation to guide the activity's flow.

With the character ready and inserted into the automaton, the child moved on to the next step of the workshop. Using the micro:bit CreateAI platform, each child trained three distinct movements for their character: fast, regular, and stop, with their gestures. In the technical setup, a micro:bit captured the children’s physical gestures and transmitted commands via BLE to an ESP32 microcontroller, which controlled a DC motor in the automata.

The experience allowed the children to perceive, in a practical and embodied way, that the automaton’s behavior depended on the examples they provided. By testing and adjusting their movements, they developed an initial understanding that automatic does not mean something that happens on its own, but rather something that can be taught, adjusted, and transformed. Instead of formalizing technical concepts, the proposal prioritized the lived experience of cause-and-effect relationships between human action and machine response.

The results indicate that the contextualized integration of AI in Early Childhood Education is feasible when anchored in concrete, playful, and embodied experiences. The children formulated hypotheses, revised strategies, and collectively celebrated outcomes, exercising agency and authorship. AI was presented as an explorable language rather than as something magical or inexplicable, fostering an approach in which children actively participated in decision-making, tested hypotheses, negotiated interpretations, and understood technology as something open to human intervention.

By engaging with the initial question of what is automatic, the children expanded their understanding through experimentation: the automatic can be constructed, tested, and modified. The experience suggests that, when mediated by investigative practices centered on children’s agency, AI can be meaningfully integrated into early childhood experiences.