Children Read Intent in Human Eyes but Not in Robots
AI-summarised brief · reviewed before publication
A study in developmental psychology and artificial intelligence found that children as young as 3 years old can read intentions in human eyes, but not in robots. The study, led by Professor Antonella Marchetti, evaluated Italian children aged 3 to 5 years old, assessing their ability to deduce preferences from human and robotic gazes. Children interpreted human gaze as meaningful, but robotic gaze was not enough to attribute psychological preference. The findings redefine engineering standards for embodied AI and child-robot interaction, highlighting the need for richer interactions beyond isolated mimicry, with significant implications for autism spectrum support and socio-communicative rehabilitation.
💡 Why It Matters
- · Human-like interaction is crucial for building trust between children and robots, and simply mimicking human signals is not enough.
- · Embodied AI that incorporates physical gestures and shared presence is necessary to create a functional connection with children.
