Natural Language Generation (NLG) is a field dealing with automatic generation of natural languages.
Natural Language Generation (NLG) is used in systems which communicate in a natural language, such as a program to automatically build a weather report based on weather data.
NLG is often thought as the reverse of sentence parsing in Natural Language Processing (NLP), but it is not, since sentence parsing only deals with the syntactic representation of a language, while NLG deals much deeper than that, touching the semantics (meaning) of language1.
Although similar in nature, NLG is significantly more difficult than NLP, due to the inherent expressiveness of natural language. NLG involves making many choices, e.g. which content to include, what order to say it in, what words and syntactic constructions to use2.
See also the Wikipedia page on Natural Language Generation: http://en.wikipedia.org/wiki/Natural_language_generation
Linked resources: