There are two schools of thought I subscribe to that lead the development of Artificial General Intelligence. The first school is based on neuroscience. It assumes the only example of working and useful intelligence we have is the human brain. Hence we need to reverse engineer its algorithms and repurpose findings as the building blocks of a hypothetical AGI. Numenta and DeepMind are the clear leaders of this approach.
The second school places the problem of intelligence at the forefront. It asks a question of what we, as intelligent creatures, do and tries to explain how we do it with algorithms. Cognitive sciences, psychology, linguistics are the primary sources of inspiration for this approach. Biology takes the back seat as a source of unnecessary implementation constraints.