Its use in psychology, sociology, zoology, and related fields where it refers to individual cognitive processing is probably the only legitimate use. As used in computer science and plant operations it functions as a metaphor. As for insect colonies, see my post above.
The use in psychology is not independent of the use within computer science. In fact, cognitive science (often seen as a branch of psychology, and is in anycase centrally concerned with the study of the brain) is both the field within which "intelligence" is used most often and is most clearly defined. Yet, as with most of psychology in general, the use developed partially out of work within computer science. Within sociology, the term is almost completely defined by its use within psychology.
Terms like "learning", "intelligence", "awareness", "memory", and so forth have always posed difficulties for specialists because 1) they are "fuzzy" 2) they are not independent of each other or other similar concepts and 3) empirically usable (let alone empirically based) definitions are notoriously problematic.
That said, a system capable of adaption to stimuli, reaction to stimuli, problem solving, memory, learning, and so forth is usually defined within all relevant fields as in some way intelligent. Any definition within psychology or some other field you mentioned (and it isn't exactly clear why you find, for example, a field like sociology uses the term legitimately yet not computer science where a great deal of artificial intelligence research is done) which did not use intelligence to describe such systems would have to abandond the term altogether. Otherwise, biologists, psychologists, and others who study learning, memory, and cognition would be left with completely arbitrary cut-off points.
Regardless, if you regard psychology or zoology as fields within which the term is used legitimately, than it is used legitimately to describe insect colonies, plants, slugs, snails, etc.