While AI is often misunderstood, due to erroneous portrayal in Hollywood movies the reality is that for libraries and in particular for the classification of articles, images, books and collections which is normally an arduous task. The machine learning handles classification particularly well and according to this article in Edtechnology the classifier algorithm is able “to ‘learn’ to recognise patterns that are common to a particular class”. There are, however, limits to the ability of a system when data sets are gathered in terms of “what we can learn from them based on their composition”. Therefore a very broad collection of representative data is fundamental, but what is also needed is human involvement in the very design of the processes that AI will follow. At the end of the day a library and the ability to be able to search for information is not just based on the amount of data available, more about the quality of that data.