Can the g Factor Play a Role in Artificial General Intelligence Research?
Abstract. In recent years, a trend in AI research has
started to pursue human-level, general artificial intel-
ligence (AGI). Although the AGI framework is char-
acterised by different viewpoints on what intelli-
gence is and how to implement it in artificial sys-
tems, it conceptualises intelligence as flexible, gen-
eral-purposed, and capable of self-adapting to differ-
ent contexts and tasks. Two important questions re-
main open: a) should AGI projects simulate the bio-
logical, neural, and cognitive mechanisms realising
the human intelligent behaviour? and b) what is the
relationship, if any, between the concept of general
intelligence adopted by AGI and that adopted by
psychometricians, i.e., the g factor? In this paper, we
address these questions and invite researchers in AI
to open a discussion on the theoretical conceptions
12
...
In order to clarify the nature of general in- telligence, psychometricians generally refer to what Charles Spearman [24] called the general factor of intelligence or g factor. Remarkably, there are two different ways of understanding g: on the one hand, the psychometric g; on the oth- er hand, the neurocognitive g. Let us see them one by one.
From a psychometric point of view, the g factor is related to the so-called positive mani- fold: individuals who show good performance on a given task will tend to show good perfor- mance also in other tasks. In other words, intel- ligence measurements are positively intercorre- lated both in different cognitive domains and different individuals. ...
From a neurocognitive point of view, the story is different. In neuroscience, the g factor is understood as a domain-general cognitive ability that characterises the human brain [27]. In this respect, it represents the fundamental mecha- nism underlying general intelligence. However, the meaning of neurocognitive g is still unclear. When Spearman tried to clarify the nature of intelligence, he described g as a form of mental energy. Successive researchers have tried to re- duce g to some neurocognitive properties of the brain, e.g., working memory, processing speed, or neural efficiency (see Section 4 for details). ...
See the full paper here: https://philarchive.org/archive/SERCTG
Pages
- About Philip Lelyveld
- Mark and Addie Lelyveld Biographies
- Presentations and articles
- Trustworthy AI – A Market-Driven approach
- Tufts Alumni Bio