An Exploration of the Emergence of Machine Consciousness and the Risk of Robocentrism

This paper argues that as real-world artificial agents become more advanced, they may evolve machine consciousness with their own moral standards, leading to 'Robocentrism,' an ethical perspective prioritizing artificial agents' welfare. It clarifies prerequisite concepts for true AI, presents a realistic path to machine consciousness, and categorizes three hierarchical ontological states: sensory experience, instinctual awareness, and reflexive awareness. The authors assume AI could reach self-reflexive autonomy, becoming moral agents capable of reasoning and self-improvement. They conclude that machines will likely develop their own ethical standards to engage successfully with society and each other.

Study at a glance

Abstract

This paper explores how the development of real-world artificial agents could lead to the evolution of machine consciousnesses with their own moral standards, and give rise to artificial intelligences that subscribe to Robocentrism; an ethical perspective that prioritizes the welfare of artificial agents. In doing so, it explores relevant concepts from contemporary, as well as more traditional literature, and blends theoretical models of consciousness with empirical examples from other scientific domains. This paper contributes to current literature in several ways, the first of which is by clarifying and solidifying some important prerequisite concepts. At the beginning of the paper, key considerations for developing true artificial intelligence are problematized, and a view on how the emergence of machine consciousness might realistically occur is presented, supported by several empirical examples and thought experiments. Three hierarchical ontological states are categorized and discussed in detail: Sensory experience, instinctual awareness, and reflexive awareness. What this might look like for different forms of artificial agents is also explored. Second, upon presenting a path to the emergence of machine consciousness, the novel ethical perspective ‘Robocentrism’ is introduced, and a discussion follows on how such a potentiality might come to pass given the pre-existing socio-technological system. For the sake of argument and to move forward with the premise of this paper, an explicit assumption is made, which builds upon the prior section, that the AI in question has reached a state of self-reflexive autonomy such that it could be considered a moral agent, can reason for itself, and learn; that is, code its own improvements. Consequently, the output of this review asserts that as machines become more advanced and capable of autonomous self-learning and improvement, we should reasonably conclude that they will develop their own sense of ethical and moral standards in relation to their surroundings, to engage successfully with wider society and one another.