Media abundance and images proliferation have significantly altered our receptive and perceptual capacity, even fuelling alternative truths, and causing problems of meaning.At the same time, the increasing presence of artificial intelligence and computational thinking in everyday life has allowed to investigate several aspects of the Human-Machine relationship, also hypothesizing some theoretical considerations on the similarity between human and robotic logic models, as well as practical steps (also in the field of learning) of the relationship between the body and the object.The history of cinema has provided many examples of how humanity can look through mechanical eyes: in this sense, one could imagine a sort of cognisphere, i.e. a hybrid mode of computational thinking and pragmatics that today manifests itself mainly in images.The aim of this paper is to hypothesize it is possible to construct narratives through the interaction between images (and their hybrid surplus of reality/digital representation) and robotics, focusing on that robots can see and transmit to us, through sensors and minicams, what they perceive. We will also refer to a series of experiments (including educational ones) that we are carrying out in the 'L. Gallino' Laboratory of Behavioural Simulation and Educational Robotics of the Department of Philosophy and Educational Sciences of the University of Turin, Italy.