A pre-neural goal for artificial intelligence

Abstract

From its onset, the discipline of Artificial Intelligence aimed at understanding intelligence through a synthetic approach. Over time, progress has been made by considering lower and lower levels of intelligence. I argue that this trend should be completed by its next step by considering pre-neural forms of intelligence as models for AI. To justify the relevance of such primitive cognition to intelligence, I recall the works of Piaget, Jonas and Maturana and Varela. By considering how these authors relate to the question of teleology, I illustrate the kind of insights a pre-neural AI could provide, which pertain to fundamental aspects of natural cognition.