🧪 Conscious Artificial Intelligence

This simulation investigates the theoretical possibility of creating artificial intelligence systems exhibiting consciousness by building the most concrete building block researchers actually study: a working self-model. A jointed robotic arm moves under its own motor commands while an internal comparator learns, from experience alone, to predict the sensory consequence of each command it issues — the same "efference copy" mechanism proposed to explain how biological brains tell self-generated sensation from the outside world. Drag the amplitude and frequency sliders to change the arm's own motion, watch the self-model's prediction error fall as it learns, then dial in an external perturbation to see the same comparator immediately re-classify the input as non-self. It is a small, honest, and fully computational answer to how a system might build awareness and self-representation from nothing but its own sensorimotor history.