Artificial intelligence (AI) can be defined as the ability of a computer or machine to replicate the problem-solving and decision-making processes traditionally associated with the human mind. The modern concept of AI can be traced back to computing machinery and intelligence, Alan Turing’s 1950 paper that explores a machine’s ability to think. The paper included the first iteration of the eponymous Turing test, a landmark development in the early development of AI.
While the nature and various components of the Turing test have changed over time, the basics have remained largely the same. The test’s purpose is to evaluate a machine’s ability to demonstrate truly intelligent behavior, to the point that a human being cannot distinguish the machine’s thought processes from those of a fellow human.
The test involves three individuals, one of which is a machine. None of the participants can see one another, though the test subject is aware that they are engaging with two individuals and that one is a computer or machine. The test is carried out through text-based channels, such as a computer screen and keyboard. A machine is considered to have passed the Turing test if the human subject cannot confidently tell the machine apart from the other human. Despite the term artificial “intelligence,” the Turing test is less concerned with the accuracy of the machine’s responses and more so with how human the answers are in nature and delivery.
Initially called the imitation game, the Turing test primarily involves the human test subject posing a series of questions to the opposing parties. Turing’s paper posited the question, “Can machines think?” Following his initial tests, he concluded, yes, they can. The Turing test has been as influential as it is controversial, though its place in the conversation and philosophy surrounding AI development is unassailable. More than 70 years later, the Turing test is still referenced in current AI discourse, as well as popular cultures, such as HBO’s Westworld and films like Ex-Machina.
Another landmark event in the history of AI is the creation of the Deep Blue supercomputer, which was designed to play chess in 1985. The system, which sourced data from 700,000 grandmaster games, could analyze approximately 200 million positions per second.
In 1996, Deep Blue was pitted against world champion Garry Kasparov in a six-game match. Deep Blue became the first computer to win against a reigning world champion under tournament rules in the first game. That said, Kasparov won three of the following five games and drew two, for a final score of 4–2, Kasparov. Deep Blue was upgraded and won a 1997 rematch with Kasparov by a score of 3.5 to 2.5, with draws counting for half a point.
Following the match, Kasparov said the computer sometimes demonstrated considerable creativity with its moves, going so far as to suggest illegal human intervention, a charge IBM scientists firmly denied.
The Deep Blue vs. Kasparov matches furthered AI research and raised public awareness of the field. Today, parts of the supercomputer are kept on display at the National Museum of American History. A 2003 documentary film, Game Over: Kasparov and the Machine, explored the event and several books.