Computers, solely based on the Turing machine, using technologies such as genetic programming, could demonstrate intelligence via their ability to solve problems, hypothesize new theorems, or produce artistic creations; this could be thought of as looking from an outsider’s point of view, and basing the measure of intelligence of a computer on the output it produces from given input. Additionally, a variant on a Turing machine where randomization is allowed could possibly add to a computer the indeterminism needed to form intelligence.
Two other ways computers could become intelligent are from either a localized or distributed form. A localized form of intelligence could be created out of such possibilities as the use of context-switching to simulate free will, or from the running of a particular program. A distributed form could come about via the massive transfer of information that is occurring via such means as the Internet, whereby a self-consciousness would be created that exists in this context, but not on any specific computer; this would be analogous to the emergent theory of consciousness.
On a different tangent, it could be argued that human beings are fully deterministic machines themselves, and as such, do not possess free will in its own right. This would make the semantic difference between computer intelligence and human intelligence a triviality: at a certain time, computers could overtake humans, gaining the ability to process such things as visual input data faster than humans, and could thus be thought to be as intelligent as a human.
If the definition of computer could be stretched to include a combination
of other technologies such as nanotechnology and cybernetics, it is possible
that intelligence to arise by utilizing a biological or inherently molecular
base. Or, along the lines of the previous, as electronic neural implants
and cybernetic technology is infused in humans, it will become impossible
to differentiate between a computer and human being and the question will
be a moot point.