Teaching computers to understand human languages
A set of algorithms has been developed that will help teach computers to process and understand human languages. The algorithms will enable a computer to act in much the same way as a human would when encountered with an unknown word. When the computer encounters a word it doesn't recognize or understand, the algorithms mean it will look up the word in a dictionary (Such as the WordNet), and tries to guess what other words should appear with this unknown word in the text. Ref. Source 6y.
If they are able to perfect it then that could mean a whole new level of translation solutions that brings us closer to the Star Trek's version of the translator. Others can get rid of the Word grammar checker in favor of something better.
Developing an algorithm to parse human language is not easy, and even trying to have it look up any unknown words in a dictionary can be problematic. Unless it has access to many different 'dictionaries', most programs will fail, as slang is quite pervasive in speech and writing. I have seen huge strides made by large translation softwares, but I think they are still far from being able to successfully parse human language in any kind of 'natural' or usual context. I hope this technology develops soon, as I love exploring the nuances of languages. But loan words, slang, and an ever-evolving vocabulary make this a very difficult and tricky problem.
Everyone pronounces their words differently and then there is speed and tone which make a lot of difference in speech. If they can perfect something like proper communication then that could mean the beginning of a real relationship between man and machine.