The definition of term "artificial general intelligence" is "intelligence of a machine that could successfully perform any intellectual task that a human being can".

Humans obviously can perform intellectual tasks that a human being can perform, thus humans are machines that implement general intelligence.

This is a surprisingly good point. The learning power of human and animal minds is very limited, varies along a normal curve, and doesn't seem to scale with increased computational power (ie: brain size). The decision-making power of human and animal minds is... well, absolute crap, actually, but evolution was working with the constraint that decision-theoretically good decision-making requires lots and lots of compute-power, implying a need for lots and lots of calories, constraining the level of decision-making intelligence that could evolve to consume very limited food supplies.