Artificial Consciousness, Superintelligence and Ethics in Robotics: How to Get There?

Abstract

Can future robots and AI-systems have consciousness and genuinely human intelligence – or even better, superhuman intelligence? Is it possible for them to behave ethically? Here we look at these questions from the point of view of philosophy and AI, and argue that these questions are related: their answer hinges on the fulfillment of the same condition. Starting from an analysis of the concept of consciousness, we argue that the key capacity that computers and robots should possess in order to emulate human cognition and (ethical) consciousness is the capacity to learn and apply ‘coherent webs-of-theories’. We conjecture that where classic AI has been, in essence, ‘data-driven’, the greatest leap forward would be ‘theory-driven’ AI. We review prominent work in deep learning and cognitive neuroscience to back-up this claim. This paper is an attempt at synthesis between recent work in philosophy, AI and cognitive science.