Our Friends Electric
In spite of early limitations, computing power eventually allowed AI to explode, leading to an abundance of significant developments in the field.
1914: Chess Machine
Spanish engineer Leonardo Torres y Quevedo creates the first of many chess-playing machines, able to play limited endgames without human intervention.
1921: “Robot” Coined
Czech playwright Karel Capek introduces the term “robot” in his play Rossum’s Universal Robots, derived from the Czech word “robota” meaning “forced labour”.
1949: Organization of Behavior
Canadian psychologist Donald Hebb proposes a theory of learning based around neural networks in his paper, Organization of Behavior: A Neuropsychological Theory.
1950: The Turing Test
British mathematician Alan Turing publishes Computing Machinery and Intelligence, in which he proposes “the imitation game” later known as the Turing Test, and suggests, for the first time, that machines can actually think.
1951: First Neural Network
US Air Force funds cognitive scientist Marvin Minsky and physicist Dean Edmonds to build the first artificial neural network, based on Hebb’s theory of learning. Known as SNARC (Stochastic Neural Analog Reinforcement Computer), the machine uses 3,000 vacuum tubes to simulate a network of 40 neurons.
1955: “AI” Coined
The term “artificial intelligence” is collectively coined in a proposal for a “two-month, 10-man study” of the field, by representatives from Dartmouth College, Harvard University, IBM and Bell Telephone Laboratories.
German American computer scientist Joseph Weizenbaum of MIT develops a chatbot capable of holding written conversations with humans.
1974–1993: AI Winter
Reduced interest and funding of AI research and programmes leads to 30 years of stagnation in the field.
1997: Deep Blue
IBM’s Deep Blue becomes the first chess-playing program to defeat a reigning world chess champion, beating Garry Kasparov 3.5 games to 2.5.
MIT’s Cynthia Breazeal debuts Kismet, the first emotionally intelligent robot capable of responding to, and simulating human feelings.
Japanese manufacturer Sony launches the AiBO (AI robot), a consumer robot in the form of a pet dog whose skills and personality develop over time.
US manufacturer iRobot release the Roomba, an autonomous home vacuum cleaner that can learn how to navigate around homes.
2004: DARPA Grand Challenge
The inaugural DARPA Grand Challenge takes place in the Mojave Desert, a 150-mile race for autonomous vehicles. None of the entrants complete the course.
2009: Large-scale Deep Unsupervised Learning
Stanford University’s Rajat Raina, Anand Madhavan and Andrew Ng publish Large-scale Deep Unsupervised Learning Using Graphics Processors, suggesting that modern graphics processors have the potential to revolutionise AI learning methods, operating 70 times faster than multicore CPUs when engaged in unsupervised learning.
2011: Siri and Watson
Apple integrates Siri, an AI assistant with voice interface, into its iPhone 4s. The same year, IBM’s Watson, a natural language question-answering computer wins $1m on US TV quiz show Jeopardy.
Microsoft’s chatbot Tay upsets a number of people on social media with offensive racist comments and is shut down after only 16 hours. Microsoft blames trolls who attacked the service as Tay learns from its interactions with others on Twitter.
Google’s AlphaGo AI defeats world Go champion Ke Jie in three straight games in a board game known for its vast number of move combinations.
2018: AlphaGo Zero
Google’s updated version of AlphaGo teaches itself to play the world’s most complex board game in three days, without any human instruction.