Consciousness defined: requirements for biological and artificial general intelligence (2406.01648v1)

Abstract: Consciousness is notoriously hard to define with objective terms. An objective definition of consciousness is critically needed so that we might accurately understand how consciousness and resultant choice behaviour may arise in biological or artificial systems. Many theories have integrated neurobiological and psychological research to explain how consciousness might arise, but few, if any, outline what is fundamentally required to generate consciousness. To identify such requirements, I examine current theories of consciousness and corresponding scientific research to generate a new definition of consciousness from first principles. Critically, consciousness is the apparatus that provides the ability to make decisions, but it is not defined by the decision itself. As such, a definition of consciousness does not require choice behaviour or an explicit awareness of temporality despite both being well-characterised outcomes of conscious thought. Rather, requirements for consciousness include: at least some capability for perception, a memory for the storage of such perceptual information which in turn provides a framework for an imagination with which a sense of self can be capable of making decisions based on possible and desired futures. Thought experiments and observable neurological phenomena demonstrate that these components are fundamentally required of consciousness, whereby the loss of any one component removes the capability for conscious thought. Identifying these requirements provides a new definition for consciousness by which we can objectively determine consciousness in any conceivable agent, such as non-human animals and artificially intelligent systems.

• The paper defines consciousness using six core components—perception, memory, imagination, self-awareness, desire, and decision-making—to form a unified framework.
• The framework bridges gaps in traditional theories by offering objective criteria for assessing consciousness in both biological organisms and AGI systems.
• The study emphasizes integrating sensors, memory modules, generative algorithms, and self-models to advance conscious capabilities in AI research and ethical debates.

Objective Requirements for Consciousness in Biological and Artificial Systems

The paper "Consciousness defined: requirements for biological and artificial general intelligence," authored by Craig I. McKenzie, addresses the longstanding challenge of defining consciousness in objective terms. Consciousness has traditionally been difficult to encapsulate due to its inherently subjective nature and the variety of disciplines engaged in its paper, ranging from neuroscience to philosophy. This work proposes a synthesized definition based on fundamental requirements, aiming to provide a framework that can apply to both biological and artificial entities.

Theoretical Background

The literature on consciousness is extensive and diverse. Prevailing theories such as Integrated Information Theory (IIT), Global Workspace Theory (GWT), first-order and higher-order theories offer insights but fail to fully delineate the essential requirements for consciousness. These theories often describe consciousness by its outputs or correlate it with specific biological processes, limiting their general applicability—especially to AGI.

Fundamental Requirements for Consciousness

McKenzie argues that definitions centered solely around characteristics like awareness or temporal continuity are inadequate. Instead, the paper identifies six core components essential for the existence of consciousness:

1. Perception: The capability to approximate reality through sensory information. Consciousness cannot arise in the absence of some form of perception, as this forms the basis for distinguishing self from non-self.
2. Memory: The storage of perceptual information which provides a framework for contextualizing experiences and building a sense of continuity. Memory is critical for integrating past experiences to influence current and future decisions.
3. Imagination: The ability to generate internal scenarios that are not directly perceived. Imagination includes both creative and critical thinking and is essential for predicting consequences and generating novel ideas.
4. Sense of Self: The intrinsic capability to distinguish between "self" and "non-self." It is the foundation upon which decisions and desires are based.
5. Desire: Arising as contextually relevant needs, desires drive decision-making processes. They are essential for the motivation behind choices that a conscious being makes.
6. Decision-Making: The resulting process from integrating perception, memory, and imagination guided by desires. The ability to make decisions and thus enable choice behavior, even if such behavior is not immediately evident (as in dreams), is indicative of consciousness.

Implications and Applications

Biological Systems

Understanding these requirements can illuminate the consciousness of non-human animals and cognitively impaired individuals. For example, this framework could help determine whether certain animals exhibit a form of rudimentary consciousness, potentially leading to ethical considerations in their treatment.

Artificial General Intelligence

Conscious AGI remains a contentious possibility. According to the definitions and components discussed, AGI systems would need to integrate sensors for perception, efficient and dynamic memory storage, advanced algorithms for imagination, a distinct self-representation model, programmed desires, and decision-making frameworks. Modern AI systems like OpenAI’s GPT models already exhibit several of these traits but lack a fully integrated sense of desires and self, which could be the next significant leap toward achieving AGI consciousness.

Theoretical and Practical Directions

This structured definition provides a groundwork for future experimental and theoretical work in both neuroscience and artificial intelligence. Neurologically, identifying which brain structures and functions support these components can refine detection methods in human and animal subjects. In AI, this definition could drive the development of systems designed with explicit goals of achieving a form of consciousness by ensuring all fundamental requirements are met.

Conclusion

The paper presents a cogent argument for a nuanced and comprehensive definition of consciousness, extending beyond human-centric or purely descriptive terms to a more foundational framework. This approach holds promise for bridging gaps between biological and artificial systems, paving the way for advancements in both fields. Additionally, it raises significant ethical questions about the treatment and rights of conscious beings, whether biological or artificial. Future research can build upon this framework to explore the complexities of consciousness further, refining our understanding and application in diverse domains.