Although artificial general intelligence (AGI) doesn't have a concrete definition, Project A.G.INT. uses it to refer to artificial consciousness.
Answer: It's an implementation of a cognitive model. The human brain is also an implementation of a cognitive model; A biological hardware implementation, although being "natural" means it doesn't hold the artificial classification.
Let's say that this is the human brain:
The circles represent neurons, and the lines represent connections.
Below, highlighed in red, is the cognitive model that the human brain implements
Any non-natural implementation of such a model would qualify as AGI.
Note: The non-highlighted parts include things like places where sensory information is processed before entering conscious perception, and life support functionallity (such as organ control).
The human brain being a hardware implementation means that the hardware (neurons) and the cognitive model are 1 to 1, whereas an AGI running on a standard computer would be a software implementation, and so the model would be simulated on the hardware.
So take the model and simulate it on a computer.
This brings up an important note, and a popular point of confusion: AGI does not give computers the ability to think or "feel". It is instead the model that has these abilities. This holds true for both natural and artificial implementations. For simplicity, there is no cognitive difference between us and an AGI. An AGI can experience the world in the same way we do, and by common definition would be alive.
