Introduction: What Is the Matter Here?

Artificial General Intelligence (AGI), roughly speaking, refers to AI research and development in which “intelligence” is understood as a general-purpose capability, not restricted to any narrow collection of problems or domains, and including the ability to broadly generalize to fundamentally new areas [4]. The precise definition of AGI is part of the subject matter of the AGI field, and different theoretical approaches to AGI may embody different slants on the very concept of AGI. In practical terms, though, the various researchers in the field share a strong common intuition regarding the core concerns of AGI – and how it differs from the “narrow AI” that currently dominates the AI field. In the earliest days of AI research, in the middle of the last century, the objective of the field was to build “thinking machines” with capacity comparable to that of the human mind [2,6,9]. In the decades following the founding of the AI field, various attempts arose to attack the problem of human-level artificial general intelligence, such as the General Problem Solver [7] and the Fifth Generation Computer Systems [3]. These early attempts failed to reach their original goals, and in the view of most AI researchers, failed to make dramatic conceptual or practical progress toward their goals. Based on these experiences,