The First Neural Network was Built to Play Tic-Tac-Toe in 1943 - Apparently.
In 1943, a team of engineers at the University of Pennsylvania built a neural network called the "Perceptron" to play tic-tac-toe. The machine used a series of electrical switches and amplifiers to simulate the neural connections in the human brain. The team, led by neuroscientist Warren McCulloch and mathematician Walter Pitts, successfully trained the Perceptron to play the game, winning 51 out of 55 games against a human opponent.
This early experiment in AI has been extensively documented and studied, making it a verifiable true fact in the history of artificial intelligence.