Contrary to what people may think, AI has been around for decades.

The April 1984 edition of ANTIC magazine featured an ATARI computer programme, written by Joseph Hafner, that implemented a self-learning AI playing Tic-Tac-Toe.

This AI, originally devised by Donald Michie in 1961, was called MENACE, short for ‘Matchbox Educable Noughts and Crosses Engine’. MENACE used 304 matchboxes to represent every state of the game. Each matchbox contained differently coloured beads that were drawn at random to represent the move that the AI should make. The AI learned the best moves by being encouraged to replay winning moves and avoid losing moves – this technique is called reinforcement learning.

This shows that AI has been around for quite a long time; indeed, AI was founded as an academic discipline in 1956. Moreover, in 1950, Alan Turing devised a test for machine intelligence, aptly called the Turing test – that’s even before AI was officially a field of study!

Incidentally, in 2015, Matt Parker and his team recreated the original MENACE, with matchboxes, beads and all, at the Manchester Science Festival. After losing a considerable number of games, the AI learned the correct strategy and soon after became unbeatable.

A video of this event may be found here, where many details are elucidated.