In AI, we often need to represent sequences: of words in sentences; of actions in plans; of sequences of moves in chess; .... This can be done using fixed-length arrays of the type described above, but it's not at all convenient to do so. So with the growth of AI and developments in symbol-processing computing went research into languages that let you manipulate variable-length sequences or lists. One of their objectives was that you should not have to worry about how to allocate storage for new items in the list, or how to de-allocate storage for old items no longer needed (something which you still have to do in Pascal and C).
There were some early attempts at this that looked like assembler with lists built in. The first general-purpose language that let you handle lists in a natural way was Lisp, developed by John McCarthy. Lisp is still one of the main AI programming languages.