I recently stumbled across this article that traces the history of Artificial Intelligence, from the first paper using the term and describing its goals in 1958 to today. Along the way, I also found
this article that worries that all of AI so far has really just variations on one computational model (“Deep Learning”) and the limitations of that model. The good news – the brilliant researcher who invented the back-propagation technique that powers deep learning is working on expanding it to conquer its greatest limitation – failure to generalize learning across domains.