Intelligence as a Measure of Consciousness
arXiv Preprint Archive August 30, 2023 Igor Ševo
Evaluating artificial systems for signs of consciousness is increasingly pressing, and a rigorous psychometric measurement framework may be crucial for assessing large language models. Most prominent theories of consciousness argue for different kinds of information coupling as necessary for human-like consciousness. By comparing information coupling in human and animal brains, cognitive development, emergent abilities, and mental representation development to analogous phenomena in large language models, the author argues that psychometric measures of intelligence, such as the g-factor or IQ, indirectly approximate the extent of conscious experience. Based on scientific and metaphysical theories, all systems possess a degree of consciousness ascertainable psychometrically, and psychometric measures may gauge relative similarities of conscious experiences across artificial and human systems.