Izak Tait

Conscious sentient AI seems to be all but a certainty in our future, whether in fifty years’ time or only five years. When that time comes, we will be faced with entities with the potential to experience more pain and suffering than any other living entity on Earth. In this paper, we look at this potential for suffering and the reasons why we would need to create a framework for protecting artificial entities. We look to current animal welfare laws and regulations to investigate why certain animals are given legal protections, and how this can be applied to AI. We use a meta-theory of consciousness to determine what developments in AI technology are needed to bring AI to the level of animal sentience where legal arguments for their protection can be made. We finally speculate on what a future conscious AI could look like based on current technology.

The “self” does not exist within a vacuum. For an entity to be considered to have a sense of self, it requires certain characteristics and attributes. This paper investigates these “structures” of the sense of self in detail, which range from a unified consciousness to self-awareness to personal identity. The paper details how each attribute and characteristic is strictly necessary for an entity to be classified as having a self, and how the five structures detailed within may be used as a guide for categorising and classifying entities as having selfhood or not (or any point along the spectrum between these). The five structures do not represent a theory of selfhood, but rather a meta-theory on the potential emergence and classification of the self.

For consciousness to exist, an entity must have prerequisite characteristics and attributes to give rise to it. We explore these “building blocks” of consciousness in detail in this paper, which range from perceptive to computational to meta-representational characteristics of an entity’s cognitive architecture. We show how each cognitive attribute is strictly necessary for the emergence of consciousness, and how the building blocks may be used for any entity to be classified as being conscious. The list of building blocks is not limited to human or organic consciousness and may be used to classify artificial and organisational conscious entities. We further explore a list of attributes that seem intuitively necessary for consciousness, but on further investigation, are neither required nor sufficient. The building blocks do not represent a theory of consciousness but rather a meta-theory on the emergence and classification of consciousness.

The established theories and frameworks on consciousness in the academic literature as related to artificial intelligence (AI), are rooted in anthropocentricism. Even those theories created intentionally for AI are based on the levels of consciousness as it is understood in humans primarily, and in other animals secondarily. This paper will discuss why such anthropocentric frameworks are built on unsecure foundations. We will do this by comparing the capacities and functions of human and AI cognitive architectures, discussing the ramifications and consequences of the behaviors that stem from these, and looking at the neurological conditions in humans that can give the most promising hints as to what a potential conscious AI entity would look like. The paper ends with a proposed solution for building a nonanthropocentric foundation of cognition that could lead toward a truly AI-focused framework of consciousness.