I guess it's not surprising there is no single discovery that started AI. Every step forward relies on a dozen theories and discoveries that went before it, each of which was essential to that final step. But let's break down the discovery of AI into three key steps.
The first step in any new invention is the idea. Today's science fiction writers excel in imagining how AI might change the societies of the future, well before any technology arrives. For the first mentions of AI, we can look back to 1726 when Jonathan Swift penned his classic satirical work, Gulliver's Travels. In it, Swift describes “The Engine,” a device for writing books that has been described as perhaps the first description of a modern computer.
The second step is the creation of the theories that form the framework of an invention. These often precede any invention by many years, and indeed the 200 years that followed Swift's Engine were a time of incredible progress in mathematics. For AI, two theories stand pre-eminent. In 1763, Thomas Bayes created an equation that describes how one's knowledge of the world can be refined by observations of what has happened before, thereby enabling the concept that we now call 'machine learning'. And a century later in 1854, George Boole described his namesake logic that forms the language of how computers operate.
So, we move to the final step: using these theories to put the ideas into practice. For AI, that belongs to the pioneers of the 20th century. But which of them deserves the credit for inventing AI? Could it be Leonardo Torres y Quevedo's 1913 primitive chess computer? Or perhaps Simon, Newell, and Shaw's "Logic Theorist," a computer program that could prove mathematical theorems by itself. Or if AI is about doing what human beings do, then maybe Weizenbaum's 1966 program ELIZA, which could hold a conversation with a human, is the key to real AI.
I think if you asked ten AI experts the same question, you might get ten different answers. Maybe I should write a computer program to determine which of them is right...