Research Paper Proposes Quantifiable Framework for Defining and Measuring Artificial General Intelligence
By
pegasus
An everything bagel for the brain. Substantive, layered, well-seasoned.
Summary
This research paper addresses the lack of a concrete definition for Artificial General Intelligence (AGI) by introducing a quantifiable framework grounded in Cattell-Horn-Carroll theory, the most empirically validated model of human cognition. The framework defines AGI as matching the cognitive versatility and proficiency of a well-educated adult, breaking down general intelligence into ten core cognitive domains including reasoning, memory, and perception. The authors adapt established human psychometric batteries to evaluate AI systems, revealing that contemporary models like GPT-4 (27%) and GPT-5 (57%) show highly 'jagged' cognitive profiles with critical deficits in foundational cognitive machinery like long-term memory storage, despite proficiency in knowledge-intensive domains.
Key quotes
· 5 pulledThe lack of a concrete definition for Artificial General Intelligence (AGI) obscures the gap between today's specialized AI and human-level cognition.
This paper introduces a quantifiable framework to address this, defining AGI as matching the cognitive versatility and proficiency of a well-educated adult.
The framework dissects general intelligence into ten core cognitive domains-including reasoning, memory, and perception-and adapts established human psychometric batteries to evaluate AI systems.
Application of this framework reveals a highly 'jagged' cognitive profile in contemporary models.
The resulting AGI scores (e.g., GPT-4 at 27%, GPT-5 at 57%) concretely quantify both rapid progress and the substantial gap remaining before AGI.
