Tsehaye Haidemariam

All Knowledge is Not Embodied: The Case of Embodied and Non-Embodied Artificial Intelligence

In the pursuit of understanding knowledge, we find ourselves grappling with the notion of embodiment. Although it is tempting to attribute all knowledge to our bodily experiences, the rise of artificial intelligence challenges this assumption. Embodied AI (such as robots), which focuses on sensory perception and physical interactions through sensors and effectors offers fascinating insights into the role of the body and environment in cognition. However, can we truly confine knowledge to the boundaries of our corporeal existence?

Enter non-embodied AI, symbolic systems, and machine-learning algorithms. These nonphysical entities lack the sensory apparatus and tactile engagement that supposedly ground our or robots’ understanding. However, let us not dismiss them hastily. They possess a remarkable capacity to process data, recognise patterns, and exhibit intelligent behaviours. Are we to ignore the knowledge they acquire merely because it lacks the embodiment that we hold so dear?

Abstract concepts, mathematical theorems, and logical reasoning challenge the notion that knowledge can be reduced to sensory experience alone. We must acknowledge the role of language, cultural influences, and the collective data of traditions in transcending individual embodiment. Through these non-embodied systems, we witness the potential to expand our understanding of the world. Therefore, I provoke you to question the assumption that knowledge is confined solely to the body. Recall that knowledge knows no bounds. It transcends the corporeal ventures into the realm of the abstract and beckons us to embrace the unembodied.

– Tsehaye Haidemariam