Ken Ewell (expertise in Deep Semantics and Conceptual Search Algorithms) writes:
The difference is in the space of functions or operations.
An intelligent computer algorithm (e.g. a binary decision procedure) runs (the functions of the intelligence operation; i.e. outputting a 1 or 0 based on input activity) in a well-managed virtual memory space under the strict control of the programming. The knowledge and intelligence of the authors are strictly replicated in the programming. The algorithm is not free to learn anything else.
Human intelligence is programmed in many ways, (e.g. cultural, religious influence, brainwashing) though it is ordinarily taught or trained by unwitting parents and teachers. Yet, unlike machines that strictly and involuntarily transform inputs into outputs, human intelligence need not depend on teachers or parents.
The precocious and playful child can learn anything with a willful aptitude for observing, grasping and discriminating the indisputable facts and relationships that give meaning to (make relevant) what is happening out of all that is taking place.
The willful aptitude appears to be a capacity (a well-developed mind). It is not an activity like intelligence. The mind appears to be a semantic (organizing) space for not only receiving information, also recognizing and triggering specified cognitive and emotional processes in light of inferred representations.