Artificial General Intelligence is the quest to build a machine intelligence that is generally intelligent, i.e., has open-ended problem solving ability and can solve problems the programmers never considered.
Specifically, the term denotes a machine of general problem-solving ability roughly equal to or greater than human beings. The term was popularized by AI researcher Ben Goertzel sometime around 2002 . Later, it became the subject of an edited volume .
Routes to AGIEdit
Proposed routes to AGI usually fall in one of three categories: top-down designs, bottom-up, and hybrid approaches. Around the year 2000, it was uncertain which approach would gain the most traction, but today, it seems clear that top-down designs are favored. The primary bottom-up approach, involving the use of genetic algorithms, has yielded scarce returns. Brain simulation approaches are well-funded, but generally these projects seem oriented towards learning more about the brain rather than building general problem-solvers.