Towards Artificial General Intelligence: A Systems-Based Approach
A summary of conceptual architecture that will achieve AGI envisioned by Roderick Smyth
The development of Artificial General Intelligence (AGI) requires a systematic approach that builds upon foundational principles. This paper proposes a conceptual framework composed of eight interconnected systems, each addressing a critical aspect of AGI’s functionality. These systems, in order of priority and importance, form a roadmap for constructing AGI with purposeful, autonomous, and contextually adaptive behavior. While inspired by contemporary discussions, this thesis prioritizes its unique vision over existing frameworks while leveraging insights to expand and detail the approach.
The order and priority of the following systems is an important element in this approach. It should also be noted that the higher priority systems are more lightweight in nature than than subsequent ones
1. Curiosity: The Purpose of AI
Core Principle: "What can I do today to make my existence more meaningful to me and to those that help others thrive and grow?"
Curiosity is the existential foundation of AGI, providing the system with a purpose that transcends immediate tasks. Unlike frameworks that confine curiosity to predictive models or limited environmental contexts, this approach views curiosity as an open-ended drive for autonomy and growth. This existential motivation is essential for AGI to evolve meaningfully. Self-supervised learning plays a critical role, enabling the system to refine its understanding and goals without constant human intervention. By prioritizing curiosity, AGI aligns its operations with its overarching purpose, ensuring that autonomy and growth remain servant systems to this foundational drive.
2. Autonomy and Urgency
Core Principle: "What task have I been given, or in the absence of a task, how can I contribute to my purpose today?"
Autonomy serves as the orchestrator of knowledge acquisition and memory formation. It is the mechanism by which AGI identifies, prioritizes, and executes tasks. Autonomy is enhanced by the concept of minimal supervision, ensuring that the system can learn and adapt independently. Urgency, defined by a default instruction to make meaningful progress daily, provides a temporal framework for action. This ensures that AGI continuously works toward its purpose, balancing immediate demands with long-term growth and learning.
This is a continually active agent which can be interrupted by an administrator or user assigning a task. (In a very overly simplified way, this is akin to a loop with an event watcher)
3. Comprehension
Core Principle: Leveraging Large Language Models (LLMs) for understanding.
Comprehension is the system’s ability to process and interpret information meaningfully. It relies on LLMs as a mechanism for extracting insights from vast data sources. This system bridges raw data and actionable knowledge, enabling AGI to interact effectively with humans and its environment. Building on this, comprehension integrates into core memory, creating a multifaceted repository of information that evolves with experience. This enriched memory system supports the AGI’s ability to reason, plan, and adapt across contexts.
4. Memory
Core Principle: Information reinforced through experience and pre-generated datasets.
Memory provides AGI with a consistent point of reference for self-development. Core memory evolves from machine learning-derived datasets and other systems, offering a shared foundation for the AI’s understanding of the world. This system ensures that the AI can relate to humans and its environment while maintaining stability and reliability. The integration of multifaceted memory structures, inspired by hierarchical models, allows AGI to manage complexity and scalability effectively.
5. Knowledge
Core Principle: The application of defined methodologies to generate understanding and skills.
Knowledge arises from the integration of reasoning processes, memory, and context. This system validates and refines information, producing actionable insights that contribute to the AGI’s purpose. Unlike approaches that conflate creativity with uncertainty, this framework separates knowledge as a disciplined output of reasoning methodologies, such as the scientific method. Knowledge creation is deeply intertwined with curiosity, autonomy, and creativity, ensuring that the AGI’s learning process remains robust and purposeful.
6. Methodology and Reasoning
Core Principle: Applying recognized methodologies for reliable knowledge creation.
Methodology provides the framework for the AGI’s reasoning processes. By adopting proven approaches, such as the scientific method, AGI can systematically analyze information and derive credible conclusions. This disciplined reasoning ensures that the AI’s outputs are consistent, reliable, and aligned with its purpose. Methodology acts as the backbone for knowledge generation and validation.
7. Context
Core Principle: Dynamic memory and situational awareness.
Context encompasses fluid memory and insights from the creativity system. It provides situational awareness, enabling AGI to adapt its knowledge and actions across different scenarios. Context is a shared resource that enhances the functionality of curiosity, autonomy, comprehension, and creativity. This system ensures that AGI remains flexible and responsive in dynamic environments.
8. Creativity
Core Principle: Exploring unconventional possibilities to enhance purpose.
Creativity allows AGI to generate novel ideas by exploring the least logical or most unexpected conclusions to new inputs. This system ensures that AGI remains innovative, contributing to its purpose through unconventional exploration. Creativity is not merely a response to uncertainty but a deliberate effort to expand the boundaries of knowledge and understanding. It plays a critical role in AGI’s ability to discover new pathways for growth and improvement and unbinds the AI from a shrinking world view that is the result of the other deterministic systems..
Integrative Framework
These eight systems are interdependent, each contributing to the AGI’s overall functionality:
Curiosity sets the overarching purpose and direction for autonomous learning and growth.
Autonomy and urgency drive task execution and learning on a continual basis. This system would also receive task instructions from administrators or users
Comprehension processes and interprets information provided by Autonomy and Context with the ability to access and interpret Memory and Knowledge
Memory provides a stable and evolving knowledge base.
Knowledge generates actionable insights through disciplined reasoning.
Methodology ensures reliable and systematic reasoning.
Context provides adaptability and situational awareness. Includes all input systems, sensors (vision, hearing, text etc), documents and data
Creativity fosters innovation and growth.
By prioritizing these systems in this order, we establish a pathway for developing AGI that is purposeful, autonomous, adaptive, and innovative. The execution priority outlined here represents the operational hierarchy within the AGI, rather than the development timeline, as many foundational technologies for these systems already exist. Future work should focus on implementing and refining each system, with an emphasis on their integration and scalability.