Timeline of artificial intelligence

From Timelines
Jump to: navigation, search

This is a timeline of artificial intelligence, which refers to the development and implementation of computer systems or machines that can perform tasks that typically require human intelligence.

Sample questions

The following are some interesting questions that can be answered by reading this timeline:

Big picture

Summary by year

Time period Development summary More details
1940s-1950s Early work This period sees the first explorations of AI, including the development of artificial neurons, learning rules for adjusting neuron connections, and the concept of connectionism.[1][2] Expert systems, which are a type of AI, are first introduced in the early 1950s. Allen Newell and Herbert A. Simon create the first artificial intelligence program. In 1956, the term "Artificial Intelligence" is first adopted.[1] Many consider John Von Neumann and Alan Turing to be the founding fathers of the technology behind AI. They pioneer the transition from 19th century decimal logic to binary logic in computer architecture. This transition leads to the development of modern computers and their ability to execute programs based on Boolean algebra. They also demonstrate that computers are universal machines capable of performing a wide range of tasks based on programming.[3] By the 1950s, a group of scientists, mathematicians, and philosophers already become familiar with the concept of artificial intelligence (AI).[4]
1960s-1970s Knowledge-based AI During this time, AI researchers focus on developing rule-based systems that can reason and make decisions based on knowledge representations. Around this period, AI experiences significant growth.The availability and affordability of computers increase, allowing for more data storage and faster processing. Additionally, machine learning algorithms improve and people become more knowledgeable about which algorithm to use for specific problems.
1974–1980 AI winter After criticism of the lack of progress in artificial intelligence (AI), government funding and interest in the field decrease during this period. Research efforts focuse on neural networks, but progress is limited, and functional programs can only handle simple problems. AI researchers have been overly optimistic in setting their goals and have made naive assumptions about the challenges they would face. When they failed to deliver promised results, funding was cut. [5][2]
1980–1987 A boom of AI Following the period of AI winter, the field of artificial intelligence makes a comeback with the introduction of expert systems. These systems are designed to mimic the decision-making abilities of a human expert through programming. [1] AI is reignited by two sources: an expansion of the algorithmic toolkit, and a boost of funds. John Hopfield and David Rumelhart popularize “deep learning” techniques which allow computers to learn using experience.[4] Funding from the United States and Britain resume to compete with Japan's "fifth generation" computer project and its goal of becoming the global leader in computer technology.[5][2][6]
1987–1993 Second AI winter Investors and governments stop funding AI research due to high costs and inefficient results, leading to another major AI winter. This coincides with the decline of early general-purpose computers and reduced government funding. Expert systems such as XCON are cost-effective but become too expensive to maintain compared to desktop computers. At the same time, DARPA concludes that AI would not be the next big thing and redirects funds to other projects. However, by the end of the 1980s, over half of the Fortune 500 companies were involved in either developing or maintaining expert systems.[1][5][2][6]
1993–2011 Emergence of intelligent agents AI research shifts its focus to intelligent agents which are used for news retrieval, online shopping, and web browsing. Despite a lack of government funding and hype, AI thrives during the 1990s and 2000s, achieving many landmark goals. Neural networks become financially successful in the 1990s when used for optical character and speech pattern recognition.[2] Major advancements are made in all areas of AI, with significant demonstrations in machine learning, natural language understanding, vision, and other fields.[7]
2011-onward Massive data and new computing power. "Deep learning, big data and artificial general intelligence" In 2011, IBM's Watson wins Jeopardy, showcasing its ability to understand natural language and solve complex questions quickly. The AI field experiences a new boom in the early 2010s due to the availability of massive amounts of data and the discovery of the high efficiency of computer graphics card processors in accelerating learning algorithms. These advancements enable significant progress at a lower financial cost.[1][3]

Summary by country

Full timeline

Year Event type Details Country/location
4th century B.C. Greek philosopher Aristotle invents syllogistic logic, the first formal deductive reasoning system.[8]
1 AC Greek mathematician and engineer Hero of Alexandria creates automatons that operate with mechanical mechanisms powered by water and steam.[9]
1206 Ebru İz Bin Rezzaz Al Jezeri, who some consider a pioneer in cybernetic science, creates water-operated automatic controlled machines.[9]
1308 Catalan poet Ramon Llull publishes "Ars generalis ultima" (The Ultimate General Art). This work improves his method of using mechanical tools made of paper to generate new ideas by combining different concepts.[10]
1623 German professor Wilhelm Schickard invents a calculating machine capable of four operations.[11][9] Germany
1642 Blaise Pascal creates the first mechanical digital calculating machine.[8]
1666 German polymath Gottfried Leibniz releases his work Dissertatio de arte combinatoria (On the Combinatorial Art). In this work, he follows Ramon Llull's idea of suggesting an alphabet of human thought and argues that all ideas are merely combinations of a small number of simple concepts.[10]
1672 Gottfried Leibniz in Paris develops a binary counting system that forms the abstract basis of modern computers.[12][9] France
1703 Gottfried Leibniz has a foresight of how binary arithmetic could be suitable for automatic calculation.[12]
1726 Jonathan Swift releases Gulliver's Travels, a book containing a portrayal of the Engine, a contraption situated on the island of Laputa that satirizes Llull's concepts. The Engine is described as "a Project for improving speculative Knowledge by practical and mechanical Operations." According to the depiction, using this device, even an uneducated individual could produce books on various subjects, such as Philosophy, Poetry, Politicks, Law, Mathematicks, and Theology, with minimal assistance from creativity or education, but with some physical effort and at a reasonable cost. [10]
1763 English statistician Thomas Bayes develops a framework for reasoning about the probability of events. The Bayesian inference would become a leading approach in machine learning.[10][13] United Kingdom (Kingdom of Great Britain)
1801 Joseph-Marie Jacquard invents the Jacquard loom, the first programmable machine, with instructions on punched cards.[8]
1854 Self-taught English mathematician, philosopher, and logician George Boole claims that logical reasoning can be systematically carried out, similar to solving a system of equations. He develops a binary algebra that represents some "laws of thought," which is published in his work titled The Laws of Thought (1854).[8]
1863 English novelist Samuel Butler suggests that Darwinian evolution also applies to machines, and speculates that they will one day become conscious and eventually supplant humanity.[14] United Kingdom
1879 German philosopher, logician, and mathematician Gottlob Frege develops modern propositional logic in his work Begriffsschrift. This would be later later clarified and expanded by Russell,Tarski, Godel, Church and others.[8] Germany
1898 Nikola Tesla showcases the world's first remote-controlled boat at an electrical exhibition in the newly built Madison Square Garden. Tesla referred to the vessel as having "a borrowed mind."[10]
1910 Principia Mathematica is published by Bertrand Russell and Alfred North Whitehead. This book would have a significant impact on formal logic. Russell, along with Ludwig Wittgenstein and Rudolf Carnap, would pave the way for a logical analysis of knowledge in philosophy.[8] United Kingdom
1912 Torres y Quevedo constructs a chess machine called the "Ajedrecista" that utilizes electromagnets located beneath the board to play out the endgame scenario of a rook and king against a single king. This creation is believed to be the earliest example of a computer game. [8]
1914 Leonardo Torres y Quevedo, a Spanish engineer, presents a chess-playing device that can play endgames with just a king and rook against a king without any human involvement.[10]
1921 The term "robot" is first introduced by Czech writer Karel Čapek in his play R.U.R. (Rossum's Universal Robots). The word is derived from "robota," which means "work" in Czech. The play explores the idea of artificial workers who ultimately turn against their human creators.[10]
1925 U.S. electrical engineer Francis P. Houdina demonstrates a radio-controlled car called the "American Wonder" on the streets of New York City. The car is able to travel at speeds of up to 20 mph, and it could turn corners and stop on command. The car is also able to avoid obstacles, such as pedestrians and other cars. The demonstration generates a lot of interest in the concept of driverless cars. However, the technology is not yet advanced enough to make driverless cars practical, and the American Wonder would be never put into production.[10] United States
1929 The first robot ever built in Japan is designed by Makoto Nishimura and named Gakutensoku, which means "learning from the laws of nature." This robot has the ability to alter its facial expression and move its head and hands, which is accomplished through an air pressure mechanism. [10]
1931 Kurt Gödel introduces the theory of deficiency, which is called by his own name.[9] "In 1931, Goedel layed the foundations of Theoretical Computer Science and AI"[15]
1936 Konrad Zuse creates a computer with programmable capabilities called Z1, which has a memory capacity of 64K.[9]
1936–1937 English mathematician Alan Turing proposes the universal Turing machine.[8] United Kingdom
1943 Warren McCulloch, a neurophysiologist at the University of Illinois, and Walter Pitts, a mathematician at the University of Chicago, release a significant publication regarding neural networks and automatons. They suggest that each neuron in the brain functions as a basic digital processor and that the entire brain is a type of computerized machine. This concept would have a significant impact on the field of artificial intelligence and would provide a theoretical foundation for the use of neural networks in modern technology.[1][16]
1943 Concept development Arturo Rosenblueth, Norbert Wiener and Julian Bigelow coin the term "cybernetics" in a paper. Wiener would publish a popular book by that name in 1948.[8]
1943 Emil Post demonstrates that production systems are a universal computational mechanism. His work on completeness, inconsistency, and proof theory is also significant. Chapter 2 of the book "Rule Based Expert Systems" discusses the applications of production systems in artificial intelligence.[8]
1945 Literature Hungarian American mathematician George Polya publishes his best-selling book on thinking heuristically, How to Solve It. This book introduces the term 'heuristic' into modern thinking and would influence many AI scientists.[8] United States
1945 Literature American engineer Vannevar Bush publishes As We May Think, a prescient vision of the future in which computers assist humans in many activities.[8] United States
1946 The first computer, ENIAC (Electronic Numerical Integrator and Computer), becomes operational. It is so large that it occupies an entire room and weights 30 tons.[9]
1948 Hungarian-American polymath John von Neumann introduces the idea of self-replicating program.[9] United States
1949 American computer scientist Edmund Berkeley writes a book titled Giant Brains: Or Machines That Think, where he discusses the emergence of news about large machines with the ability to handle vast amounts of information at a great speed and with great skill. According to him, these machines are comparable to a brain made of wires and hardware instead of flesh and nerves. In his opinion, machines are capable of thinking because they are capable of performing logical operations, making conclusions, and decisions based on information.[10] United States
1949 Donald Hebb publishes a book called "Organization of Behavior: A Neuropsychological Theory," which proposes a theory about learning based on the ability of synapses to strengthen or weaken over time in neural networks. Hebb demonstrates an updating rule for modifying the connection strength between neurons, which would be later known as Hebbian learning.[10][1]
1950 In an article for Scientific American, Claude Shannon argues that only an artificial intelligence program could play computer chess at a high level. He points out that the number of possible moves in a chess game is so vast that it would be impossible for a human to consider all of them. An AI program, on the other hand, could use a search algorithm to explore all of the possible moves and select the best one. Shannon's article would become a landmark in the history of computer chess. It would help to lay the foundation for the development of the first chess-playing programs, which would be developed in the 1950s and 1960s. Today, AI programs are able to play chess at a level that is far superior to any human player.[10][8][17]
1950 Concept development Alan Turing publishes his article "Computing Machinery and Intelligence", which introduces the concept of the Turing Test, also known as the imitation game. This game involves a human judge trying to distinguish between a human and a machine in a teletype conversation. Turing's article is the first to raise the question of whether a machine could exhibit intelligence.[3]
1951 Marvin Minsky and Dean Edmunds build SNARC (Stochastic Neural Analog Reinforcement Calculator), the first artificial neural network, using 3000 vacuum tubes to simulate a network of 40 neurons.[10]
1951 The first artificial intelligence programs for the Harvard Mark I device are written.[9] United States
1952 American computer scientist Arthur Samuel develops the first computer checkers-playing program and the first computer program to learn on its own.[10] United States
1952 Alan Hodgkin and Andrew Huxley publish a paper in the journal Nature that describe a mathematical model of the electrical activity of neurons. The model, which would be later known as the Hodgkin-Huxley model, is a set of nonlinear differential equations that describe how the membrane potential of a neuron changes over time. The Hodgkin-Huxley model would become a major breakthrough in the field of neuroscience, and it would help to lay the foundation for our understanding of how neurons work. The model would be used to study a wide range of phenomena in neuroscience, including the generation of action potentials, the propagation of action potentials, and the integration of synaptic inputs. The Hodgkin-Huxley model is a simplified model of the neuron, but it is still a very powerful tool for understanding how neurons work.[2]
1953 Arthur Prior, a philosopher at the University of Canterbury, first introduces tense logic, which would be used by languages to express time-dependent data. Tense logic helps in locating statements in the flow of time.[16]
1954 The Georgetown-IBM experiment becomes the first demonstration of machine translation (MT). The experiment is conducted by a team of researchers from Georgetown University and IBM. They use a computer called the IBM 701 to translate 60 Russian sentences into English. The sentences are all related to organic chemistry, and the translation system was able to translate them with an accuracy of 85%. The Georgetown-IBM experiment becomes a major milestone in the history of MT. It shows that MT was a real possibility, and it paves the way for the development of more advanced MT systems.[6] United States
1954 Belmont Farley and Wesley Clark of MIT achieve a significant milestone by running the first artificial neural network. Although limited by computer memory to 128 neurons, they are able to train the network to recognize simple patterns. They also discover that damaging up to 10 percent of the neurons did not affect the network's performance, similar to the brain's ability to tolerate limited damage. The depicted neural network exemplifies the fundamental concepts of connectionism.[16]
1955 In August 31, 1955, a proposal titled 2 month, 10 man study of artificial intelligence is submitted by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon. This proposal introduces the term "artificial intelligence." The workshop, held in July and August 1956, would be widely regarded as the official birth of the field of artificial intelligence.[10]
1955 (December) Herbert Simon and Allen Newell introduce the Logic Theorist, recognized as the first artificial intelligence program. This program achieves a remarkable feat by proving 38 out of the initial 52 theorems found in Whitehead and Russell's Principia Mathematica. Additionally, it discovers new and more elegant proofs for some of these theorems.[10][1]
1955–1956 Allen Newell, J. Clifford Shaw, and Herbert Simon create the Logic Theorist, a groundbreaking program aimed at proving theorems from Principia Mathematica by Whitehead and Russell. The Logic Theorist, as it comes to be known, is capable of producing more elegant proofs than those found in the original books, marking a significant achievement in this field.[16]
1956 The inaugural "Artificial Intelligence" conference takes place at Dartmouth College in Hanover, New Hampshire. The term "artificial intelligence" was previously coined in a proposal submitted by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon in August 1955, leading to the official birth of the field during the workshop held in July and August 1956. This summer conference, funded by the Rockefeller Institute, is considered the foundation of the discipline. Remarkably, it is a workshop rather than a conventional conference, with only six participants, including McCarthy and Minsky, who would remain consistently engaged in developing the field, primarily through formal logic.[18][10] [3][6][1] United States
1956 Newell and Simon develop the Logic Theorist program, an early AI system designed to discover proofs in propositional logic. This marks the inception of artificial intelligence as a field, with Logic Theorist being later often considered the first AI program. It is presented at the Dartmouth Summer Research Project on Artificial Intelligence (DSRPAI) in the same year, a conference hosted by John McCarthy and Marvin Minsky, where the term "artificial intelligence" is first coined. The program aims to simulate human problem-solving skills and was funded by the RAND Corporation.[19][4][5][9]
1957 Frank Rosenblatt creates the Perceptron, one of the initial artificial neural networks that facilitates pattern recognition through a two-layer computer learning system. The New York Times describes the Perceptron as the early stages of an electronic computer that the Navy anticipates can eventually possess capabilities such as walking, talking, seeing, writing, self-replicating, and self-awareness. The New Yorker characterizes it as an extraordinary machine with the potential for what can be considered as thought processes.[10]
1957 Herbert Simon, an economist and sociologist, predicts that artificial intelligence would be able to defeat a human at chess within a decade. However, AI research would experience a setback and would go through a period of dormancy. Nevertheless, Simon's prediction ultimately would come true, but it would take 30 years for AI to accomplish this feat.[3]
1957 Herbert Newell, Cliff Shaw, and Herbert Simon demonstrate the General Problem Solver (GPS). This program, developed over about a decade, is capable of solving a wide range of puzzles through a trial-and-error approach, showcasing significant problem-solving abilities.[16][8]
1958 American computer scientist John McCarthy develops the Lisp programming language. Lisp is a functional programming language that is well-suited for artificial intelligence applications. It is a recursive language, which means that it can be used to represent recursive data structures, such as lists. This makes it a powerful tool for representing the knowledge that is used in artificial intelligence applications. Lisp would be used in a wide variety of artificial intelligence applications, including natural language processing, machine learning, and robotics. It is still one of the most popular programming languages used in artificial intelligence research.[10][9]
1958 Herbert Gelernter's "geometry machine" becomes the first advanced AI program to prove geometric theorems, marking a significant milestone in artificial intelligence development.[20]
1959 Arthur Samuel coins the term "machine learning" while reporting his work on programming a computer to improve its checkers game-playing skills beyond the capabilities of its human programmer.[10]
1959 Oliver Selfridge publishes Pandemonium: A paradigm for learning, which describes a model in which computers can recognize patterns that has not been pre-specified. This work lays the foundation for pattern recognition and learning in AI.[10]
1959 John McCarthy publishes Programs with Common Sense, in which he introduces the concept of the "Advice Taker," a program designed for problem-solving and common-sense reasoning.[10]
1959 Samuel creates a checkers program. Later in the late 1950s, he would design a program that can learn how to play checkers.[19]
1960 American psychologist and computer scientist J. C. R. Licklider described the human-machine relationship in his work.[9] United States
1961 Unimate, the first industrial robot, starts working on an assembly line in a General Motors plant in New Jersey.[10][21] United States
1961 James Slagle in his PhD dissertation writes in Lisp the first symbolic integration program, SAINT, which solves calculus problems at the college freshman level.[8]
1961 American computer scientist James Robert Slagle develops SAINT (Symbolic Automatic INTegrator), a heuristic program designed to solve symbolic integration problems typically found in freshman calculus.[10] United States
1963 Reed C. Lawlor, a member of the California Bar, authors an article titled What Computers Can Do: Analysis and Prediction of Judicial Decisions. The article explores the potential of computers in analyzing and predicting judicial decisions.[3]
1963 Thomas Evans develops a program called ANALOGY as part of his MIT PhD work. This program demonstrates that computers are capable of solving analogy problems similar to those found on IQ tests.[8]
1963 Ivan Sutherland's MIT dissertation on Sketchpad introduces the concept of interactive graphics into the field of computing.[8] United States
1963 Edward A. Feigenbaum and Julian Feldman publish Computers and Thought, which is the first collection of articles focused on artificial intelligence.[8]
1964 Daniel Bobrow completes his MIT PhD dissertation titled Natural Language Input for a Computer Problem Solving System and creates STUDENT, a computer program for natural language understanding.[10]
1964 The Society for the Study of Artificial Intelligence and the Simulation of Behaviour is founded. It is the oldest AI society in the world. United Kingdom
1964 Danny Bobrow's MIT dissertation demonstrates that computers can understand natural language well enough to correctly solve algebra word problems.[8]
1964 Bert Raphael's MIT dissertation on the SIR program showcases the effectiveness of a logical knowledge representation for question-answering systems.[8]
1965 Herbert Simon predicts in The Shape of Automation for Men and Management that machines would be capable of doing any work a man can do within 20 years.[22] "Herbert Simon predicts that "machines will be capable, within twenty years, of doing any work a man can do.""[10]
1965 American philosopher Hubert Dreyfus publishes Alchemy and AI, which argues that the mind is not like a computer and that there are limits beyond which artificial intelligence would not progress.[10] United States
1965 I.J. Good writes in "Speculations Concerning the First Ultraintelligent Machine" that the first ultraintelligent machine could potentially be humanity's last invention, as long as it remains compliant enough to guide us in controlling it.[10]
1965 Joseph Weizenbaum creates ELIZA, an interactive software that engages in conversations in English about various subjects. Weizenbaum's objective was to exhibit the superficial nature of communication between humans and machines. However, he would be taken aback by the number of individuals attributing human-like emotions to the computer program.[10]
1965 Edward Feigenbaum, Bruce G. Buchanan, Joshua Lederberg, and Carl Djerassi begin developing DENDRAL at Stanford University. DENDRAL is the first expert system, designed to automate the decision-making and problem-solving tasks performed by organic chemists. Its primary goal is to explore hypothesis formation and the creation of models for empirical induction in scientific research.[10][3] United States
1965 Literature Hubert Dreyfus publishes Alchemy and AI.
1965 J. Alan Robinson develops the Resolution Method, a mechanical proof procedure that enables programs to efficiently work with formal logic as a representation language.[8]
1965 Joseph Weizenbaum, a researcher at MIT, develops ELIZA, an interactive software that engages in conversations in the English language on various subjects. Initially, it is a well-liked application at AI centers on the ARPA-net. However, a modified version would be created to imitate the conversation style of a psychotherapist.[8]
1966 Shakey the robot is introduced as the first general-purpose mobile robot capable of reasoning about its own actions. An article in Life magazine in 1970 refers to Shakey as the "first electronic person," and Marvin Minsky predicts that within three to eight years, a machine with the general intelligence of an average human would be achieved.[10]
1966 Joseph Weizenbaum, a German-American computer scientist at MIT, creates the first chatbot named ELIZA. ELIZA uses scripts to simulate conversations with humans, including the role of a psychotherapist. This development highlights the early focus on algorithm development for mathematical problem-solving.[23][1]
1966 The ALPAC report, known for its skepticism about machine translation research and its call for increased focus on basic computational linguistics research, results in a significant reduction in U.S. government funding for this field. This report, along with the 1973 Lighthill report for the British government, contribute to the onset of the AI winter, a period marked by reduced funding and interest in artificial intelligence research.[6][8]
1966 Organization Canadian engineer Charles Rosen founds the Artificial Intelligence Center.[24]
1966 Ross Quillian in his PhD dissertation at Carnegie Institute of Technology demonstrates semantic networks[8], which are basically graphic depictions of knowledge composed of nodes and links that show hierarchical relationships between objects.[25] Semantic networks are an alternative to first-order logic as a form of knowledge representation.[26] United States
1966 The first Machine Intelligence workshop takes place in Edinburgh, marking the beginning of an influential annual series of workshops organized by Donald Michie and others.[8] United Kingdom
1967 The Dendral program, developed by Edward Feigenbaum, Joshua Lederberg, Bruce Buchanan, and Georgia Sutherland at Stanford University, successfully demonstrates the interpretation of mass spectra on organic chemical compounds. This achievement marks the first successful knowledge-based program for scientific reasoning.[8]
1967 Joel Moses, during his PhD work at MIT, demonstrates the effectiveness of symbolic reasoning for integration problems through the Macsyma program. This marks a significant milestone as the first successful knowledge-based program in mathematics.[8]
1967 Richard Greenblatt at MIT develops MacHack, a knowledge-based chess-playing program that achieved a class-C rating in tournament play. This achievement marks a notable advancement in computer chess.[8]
1967 Daniel Bobrow's STUDENT program demonstrates the ability to solve high school algebra problems expressed in natural language, showcasing early advancements in natural language understanding by computers.[19]
1968 Stanley Kubrick's film 2001: A Space Odyssey is released, featuring HAL 9000, a sentient computer that raises questions about the sophistication, benefits, and dangers of AI. While not a scientific contribution, the film would play a significant role in popularizing AI themes and ethical questions. Science fiction authors like Philip K. Dick also explore the idea of machines experiencing emotions, contributing to the discourse around AI. [10][3]
1968 American computer scientist Terry Winograd creates SHRDLU, an early computer program capable of understanding natural language.[10] United States
1969 Arthur E. Bryson and Yu-Chi Ho describe backpropagation as a multi-stage dynamic system optimization method. While it doesn't gain prominence immediately, this learning algorithm for multi-layer artificial neural networks would later play a significant role in the success of deep learning during the 2000s and 2010s, as computing power advances to enable the training of large neural networks.[10]
1969 Marvin Minsky and Seymour Papert publish Perceptrons: An Introduction to Computational Geometry, which highlights the limitations of simple neural networks called perceptrons. An expanded edition in 1988 would clarify that their conclusions in 1969 didn't significantly reduce funding for neural network research. Instead, they would argued that progress has stalled due to a lack of adequate basic theories in the mid-1960s, despite many experiments with perceptrons. The book emphasizes the need for a deeper understanding of why certain patterns could be recognized by neural networks while others could not.[10]
1969 Conference The first International Joint Conference on Artificial Intelligence (IJCAI) is held in Washington, D.C.[8] United States
1969 The SRI robot Shakey demonstrates the ability to combine locomotion, perception, and problem solving. This is a major breakthrough in the field of robotics, as it shows that it is possible to build a robot that can interact with its environment in a meaningful way. Shakey is equipped with a variety of sensors, including a television camera, a laser range finder, and a bump sensor. These sensors allow Shakey to see its surroundings, measure the distance to objects, and detect obstacles. Shakey is also equipped with a problem-solving system that allows it to plan its movements and solve simple problems. Shakey's success shows that it is possible to build a robot that can combine locomotion, perception, and problem solving. This is a major breakthrough in the field of robotics, as it paves the way for the development of more advanced mobile robots.[8] United States
1969 Roger Schank, a researcher at Stanford University, introduces the conceptual dependency model for natural language understanding. This model would be further developed for applications in story understanding by Robert Wilensky and Wendy Lehnert during their PhD dissertations at Yale University. Additionally, Janet Kolodner would expand its use in understanding memory.[8] United States
1970 Literature Journal Artificial Intelligence is first published by Elsevier.[27] Netherlands
1970 Waseda University in Japan creates the WABOT-1, the first anthropomorphic robot. This robot features limb control, a vision system, and a conversation system, marking a significant advancement in robotics.[10]
1970 Marvin Minsky expresses optimism to Life Magazine, suggesting that within three to eight years, a machine with the general intelligence of an average human being would be developed. However, despite the progress made in basic principles, there is still a considerable distance to cover before achieving goals like natural language processing, abstract thinking, and self-recognition in AI.[4]
1970 Uruguayan American Jaime Carbonell develops SCHOLAR, an interactive program for computer-aided instruction based on semantic nets as the representation of knowledge.[8] SCHOLAR is perhaps the first intelligent tutoring system.[28] United States
1970 Bill Woods describes Augmented Transition Networks (ATN) as a representation for natural language understanding.[8] The ATN is a formalism for writing parsing grammars that would be much used in artificial intelligence and computational linguistics.[29]
1970 Patrick Winston's PhD program called ARCH, which is conducted at MIT, focuses on teaching computers to learn concepts from examples in the context of children's building blocks.[8]
1971 Terry Winograd's MIT PhD thesis showcases computers' capacity to comprehend English sentences within a limited context involving children's building blocks. He achieves this by integrating his language comprehension program, SHRDLU, with a robot arm that executes instructions provided in English text.[8]
1972 Expert system Stanford University introduces MYCIN, one of the early expert systems designed for diagnosing severe infections, identifying bacteria responsible, and recommending suitable antibiotics. MYCIN represents a pioneering application of artificial intelligence in the medical field, serving as an expert system that utilized rules, formulas, and a knowledge database to assist in diagnosing and treating illnesses.[4][3][23]
1972 The WABOT-1 becomes the first full-scale humanoid intelligent robot built in the world. It is developed by a team of researchers at Waseda University in Tokyo, Japan, led by Ichiro Kato. The WABOT-1 is able to walk, talk, and interact with people in a limited way. A major breakthrough in the field of robotics it shows that it is possible to build a robot that could interact with humans in a meaningful way. The research that is done on the WABOT-1 would help to pave the way for the development of more advanced humanoid robots, such as the ASIMO robot developed by Honda.[1] Japan
1972 Lierature Hubert Dreyfus publishes What Computers Can't Do.[30]
1972 French computer scientist Alain Colmerauer develops Prolog, a programming language commonly used for artificial intelligence and symbolic reasoning.[8]
1972 Work commences on MYCIN, an expert system designed to diagnose blood infections. Developed at Stanford University, MYCIN aims to diagnose patients by analyzing their reported symptoms and medical test results.[16]
1972 Alan Kay, Dan Ingalls, and Adele Goldberg at Xerox PARC introduce the Smalltalk programming language. Smalltalk is a groundbreaking, purely object-oriented language primarily created for teaching programming to young individuals. It emphasizes the message-passing paradigm, marking a significant development in object-oriented programming and icon-oriented interfaces.[8][31]
1973 James Lighthill is commissioned by the head of the British Science Research Council, Brian Flowers, to evaluate requests for support in AI research. His report, "Artificial Intelligence: A General Survey," published in 1973, concludes that the discoveries made in the field of AI research had not lived up to the earlier promises of major impact. This pessimistic prognosis by Lighthill would result in reduced government funding for AI research, and his report would be commonly referred to as the "Lighthill report."[4][6]
1973 Alain Colmerauer at the University of Aix-Marseille, France, conceive the logic programming language PROLOG (Programmation en Logique), which is first implemented that same year. PROLOG would be further developed by Robert Kowalski, a logician at the University of Edinburgh. This language employs a powerful theorem-proving technique called resolution, which was invented in 1963 by British logician Alan Robinson. PROLOG is capable of determining the logical validity of statements, making it widely used in AI research, particularly in Europe and Japan.[16]
1973 DARPA initiates the development of protocols known as TCP/IP.[9]
1974 Conference European Conference on Artificial Intelligence[32]
1974 Ted Shortliffe's PhD dissertation at Stanford University showcases the effectiveness of rule-based systems in the realm of medical diagnosis and treatment, specifically focusing on MYCIN. This work is often regarded as a pioneering example of an expert system in the field of artificial intelligence.[8]
1974 Earl Sacerdoti made significant advancements in the field of artificial intelligence by developing one of the earliest planning programs known as ABSTRIPS. His work also introduced techniques for hierarchical planning, which had a substantial impact on AI planning systems.[8]
1975 Marvin Minsky publishes a highly influential article on Frames as a knowledge representation. This work brings together various ideas related to schemas and semantic links, contributing significantly to the field of artificial intelligence and knowledge representation.[8]
1975 The Meta-Dendral learning program achieves a significant milestone by generating new findings in chemistry, specifically in the realm of mass spectrometry. These results mark the first instance of scientific discoveries made by a computer that are published in a peer-reviewed journal.[8]
1976 Computer scientist Raj Reddy publishes a seminal paper titled Speech Recognition by Machine: A Review in the Proceedings of the IEEE. This paper provides a comprehensive overview of the early developments in Natural Language Processing (NLP) and speech recognition by machines.[10]
1976 AI research faces challenges as processing power fails to match the promising theoretical advancements made by computer scientists. Roboticist Hans Moravec asserts that computers are "still millions of times too weak to exhibit intelligence," highlighting the limitations in computational capabilities during that era.[33]
1976 Doug Lenat's AM program, which is the subject of his Stanford PhD dissertation, showcases the discovery model, involving a loosely-guided search for intriguing conjectures.[8]
1976 Randall Davis demonstrates the significance of meta-level reasoning through his PhD dissertation at Stanford University.[8]
Mid1970s American computer scientist Barbara J. Grosz at SRI sets limits to traditional AI approaches in discourse modeling. Her subsequent work, along with Bonnie Webber and Candace Sidner, introduces the concept of "centering," which would become important in determining discourse focus and managing anaphoric references in Natural Language Processing (NLP).[8] United States
Mid1970s British neuroscientist David Marr and his colleagues at MIT propose a theory of visual perception that includes the concept of the "primal sketch." The primal sketch is a low-level representation of the visual world that is based on the edges and textures of surfaces. It is the first step in Marr's theory of visual perception, which is a hierarchical model that describes how the brain processes visual information.[8]
1977 iLabs[34] Italy
1978 Expert system The XCON (eXpert CONfigurer) program, which is a rule-based expert system, is developed at Carnegie Mellon University. XCON aims to assist in the ordering of DEC's VAX computers by automatically selecting the components based on the customer's specific requirements. This marks an important milestone in the development of expert systems, showcasing their ability to automate complex decision-making processes.[10]
1978 Japan's Ministry of International Trade and Industry (MITI) initiates a study to explore the future of computers. Three years later, MITI would embark on a project to develop fifth-generation computers, aiming to achieve a significant advancement in computer technology. These new computers are intended to surpass existing technology, relying on multiprocessor machines specialized in logic programming instead of standard microprocessors. The goal is to position Japan as a technological leader in information processing and artificial intelligence, betting on high-power logic machines to catalyze these advancements.[6]
1978 Herbert Simon is awarded the Nobel Prize for his pioneering work on the Limited Rationality Theory, a significant contribution to the field of Artificial Intelligence.[9][8]
1978 Tom Mitchell, based at Stanford, introduces the concept of Version Spaces, a framework for describing the search space in concept formation programs.[8]
1978 The MOLGEN program, developed by Mark Stefik and Peter Friedland at Stanford, showcases the utility of an object-oriented knowledge representation for planning gene-cloning experiments.[8]
1979 The Stanford Cart achieves the significant milestone of autonomously navigating a room filled with chairs, completing the task in approximately five hours. This accomplishment marks one of the early instances of an autonomous vehicle demonstrating its capabilities.[10]
1979 The Association for the Advancement of Artificial Intelligence is founded.[35] United States
1979 The MYCIN program, initially developed as Ted Shortliffe's Ph.D. dissertation at Stanford, is demonstrated to perform at the level of experts. Another significant development is Bill VanMelle's Ph.D. dissertation at Stanford, which showcases the generality of MYCIN's knowledge representation and reasoning style in his EMYCIN program. EMYCIN serves as a model for many commercial expert system "shells," marking a milestone in the field of artificial intelligence and expert systems.[8]
1979 Jack Myers and Harry Pople at the University of Pittsburgh develop INTERNIST, a knowledge-based medical diagnosis program that leveraged Dr. Myers' clinical expertise. This program represents a significant advancement in the application of artificial intelligence to the field of medical diagnosis.[8]
1979 Cordell Green, David Barstow, Elaine Kant, and their team at Stanford demonstrate the CHI system, which is designed for automatic programming. This system marks a notable development in the field of artificial intelligence and its applications in automating programming tasks.[8]
1979 Drew McDermott and Jon Doyle at MIT, along with John McCarthy at Stanford, begin publishing research on non-monotonic logics and formal aspects of truth maintenance. Their work in this area would contribute to advancing the understanding and development of logic-based systems in artificial intelligence.[8]
Late 1970s Stanford's SUMEX-AIM resource, led by Ed Feigenbaum and Joshua Lederberg, showcases the potential of the ARPAnet for facilitating scientific collaboration, highlighting the impact of computer networks on research and information sharing in the field of artificial intelligence and beyond.[8]
1980 Computer scientist Edward Feigenbaum plays a pivotal role in rekindling AI research by championing the development of "expert systems." These systems learn by consulting experts in a particular domain to gather responses for various situations. Once these expert responses are collected and compiled for a wide range of scenarios in that domain, the expert system can offer specialized guidance to non-experts in that field, marking a significant advancement in AI research.[33]
1980 Expert system After the AI winter period, AI experiences a resurgence with the introduction of "Expert Systems." These systems are designed to replicate the decision-making capabilities of human experts, signifying a significant revival in the field of artificial intelligence.[1] AI research experiences a resurgence with increased funding and the development of algorithmic tools, including deep learning techniques, which allow computers to learn from user experiences.[36]
1980 Waseda University in Japan develops Wabot-2, a humanoid musician robot. This robot has the ability to interact with humans, read musical scores, and play moderately complex tunes on an electronic organ.[10]
1980 Expert system Digital Equipment Corporation (DEC) implements an Expert System called XCON to assist its sales team in placing customer orders. DEC, a company selling various computer components, utilized XCON because their sales force lacked in-depth knowledge about the products they were selling. This move helps streamline the ordering process and improve customer service.[2]
1980 The American Association of Artificial Intelligence (AAAI) held its first national conference at Stanford University.[1]
1980 Lee Erman, Rick Hayes-Roth, Victor Lesser, and Raj Reddy publish the first description of the blackboard model, which serves as the framework for the HEARSAY-II speech understanding system.[8]
1980 The first National Conference of the American Association of Artificial Intelligence (AAAI) is held at Stanford University.[8]
1980 The term "strong AI" is introduced by philosopher John Searle of the University of California at Berkeley to categorize a specific area of AI research.[16]
1981 An expert system called SID (Synthesis of Integral Design) is able to design 93% of the VAX 9000 CPU logic gates. This system, consisting of 1,000 hand-written rules, completes the CPU design in just 3 hours, surpassing human experts in various aspects. For instance, it produces a faster 64-bit adder than the manually designed one and achieves a significantly lower bug rate, reducing it from approximately 1 bug per 200 gates in human-designed systems to about 1 bug per 20,000 gates in the final output of the SID system.[37]
1981 Danny Hillis designs the Connection Machine, a massively parallel architecture that significantly boosts the capabilities of artificial intelligence and computing in general. This development ultimately led to the founding of Thinking Machines, Inc.[8]
1981 The Japanese Ministry of International Trade and Industry allocates a substantial budget of $850 million for the Fifth Generation Computer project. This ambitious project aims to develop computers capable of engaging in conversations, translating languages, interpreting images, and reasoning like human beings.[10]
1981 Japan's Ministry of International Trade and Industry (MITI) commissions a study to explore the future of computers. Three years later, MITI launches the ambitious Fifth Generation Computer project with a budget of $850 million. The project aims to create a new generation of computers that would represent a significant leap in technology. These machines would not rely on standard microprocessors but would be multiprocessor systems specialized in logic programming. The goal is to propel Japan to the forefront of technology by catalyzing advancements in information processing and realizing artificial intelligence capabilities.[6] Japan
1982 European Association for Artificial Intelligence
1983 Organization The Turing Institute is founded in Glasgow, Scotland as an Artificial Intelligence laboratory. The company would undertake basic and applied research, working directly with large companies across Europe, the United States, and Japan developing software as well as providing training, consultancy and information services.[38] From 1989 onwards, the company would face financial difficulties and would close in 1994.[39] United Kingdom
1983 John Laird and Paul Rosenbloom, under the guidance of Allen Newell, complete their dissertations at Carnegie Mellon University on the SOAR project.[8]
1983 James Allen invents the later called Allen's interval algebra, the first widely used formalization of temporal events.[8][40][41] Alsocalled Allen's Interval Calculus, it is certainly the most well-known qualitative temporal calculus in artificial intelligence.[42]
1984 The film "Electric Dreams" was released, depicting a love triangle between a man, a woman, and a personal computer.[10]
1984 At the annual meeting of AAAI (American Association for Artificial Intelligence), Roger Schank and Marvin Minsky warn of the impending "AI Winter." They predict a downturn in AI investment and research funding, similar to the reduction that had occurred in the mid-1970s. This prediction would indeed materialize three years later when AI research faces a decline in support and interest.[10]
1984 The CYC project is initiated as a significant endeavor in symbolic AI. This project is launched under the sponsorship of the Microelectronics and Computer Technology Corporation, a consortium consisting of computer, semiconductor, and electronics manufacturers.[16]
1985 Harold Cohen demonstrates the autonomous drawing program called Aaron at the AAAI National Conference. Aaron, which was developed over more than a decade, showcases significant advancements in autonomous drawing capabilities.[8]
1986 A team of researchers at the Bundeswehr University Munich, Germany, led by Ernst Dickmanns, builds the first driverless car, a Mercedes-Benz van equipped with cameras and sensors that allow it to navigate empty streets at speeds of up to 55 mph. The car is able to follow the road markings, avoid obstacles, and even change lanes. This is a major milestone in the development of self-driving cars, and it shows that it is possible to build a car that could drive itself safely on public roads. The research that is done on this car would help to pave the way for the development of the self-driving cars that we see today.[10]
1986 Literature Hubert Dreyfus publishes Mind over Machine.
1986 A notable connectionist experiment at the University of California in San Diego, led by David Rumelhart and James McClelland, involves training a neural network comprising 920 artificial neurons arranged in two layers (460 neurons each) to generate past tenses for English verbs. The root forms of verbs, like "come," "look," and "sleep," are fed into the input layer. A supervisory computer program observes the output layer's response and the desired response (e.g., "came") and adjustes network connections accordingly. After approximately 400 verb presentations, repeated 200 times, the network can correctly generate past tenses for both familiar and unfamiliar verbs.[16] United States
1986 (October) Organization The Centre for Artificial Intelligence and Robotics is founded in Bangalore as a laboratory of the Defence Research & Development Organization.[43] India
1986 (October) David Rumelhart, Geoffrey Hinton, and Ronald Williams publish a groundbreaking paper titled "Learning representations by back-propagating errors." This paper introduces a novel learning procedure known as back-propagation, designed for networks of neuron-like units. Back-propagation would later become a fundamental technique in training artificial neural networks, contributing significantly to the success of deep learning in subsequent decades.[10]
1986 Terrence J. Sejnowski and Charles Rosenberg introduce the 'NETtalk' program, a significant achievement in the development of artificial intelligence. 'NETtalk' is capable of speech synthesis, which allows a computer to speak for the first time. It learns to speak by processing sample sentences and phoneme chains. Moreover, 'NETtalk' has the ability to read and correctly pronounce words, and it can apply its learning to words it has never encountered before. This program is one of the early examples of artificial neural networks, which functions similarly to the human brain and learns from extensive datasets.[23]
1986 Conference International Conference on User Modeling, Adaptation, and Personalization
1987 A video titled "Knowledge Navigator" is presented during Apple CEO John Sculley's keynote speech at Educom. This video depicts a futuristic vision in which "knowledge applications would be accessed by smart agents working over networks connected to massive amounts of digitized information."[10]
1987 Literature AI & Society
1987 Literature Applied Artificial Intelligence
1987 Literature International Journal of Pattern Recognition and Artificial Intelligence
1987 Marvin Minsky publishes "The Society of Mind," a theoretical work that describes the mind as a collection of cooperating agents.[8]
1988 Judea Pearl publishes Probabilistic Reasoning in Intelligent Systems, laying the foundation for processing information under uncertainty. His pioneering work includes the invention of Bayesian networks and algorithms for inference in these models, which revolutionized artificial intelligence and found applications in various engineering and scientific fields. He would be later awarded the Turing Award for his contributions.[10]
1988 Dalle Molle Institute for Artificial Intelligence Research Switzerland
1988 Rollo Carpenter develops Jabberwacky, a chat-bot aimed at simulating natural human chat in an entertaining and humorous manner. This marks an early attempt at using human interaction to create artificial intelligence.[10]
1988 Members of the IBM T.J. Watson Research Center publish a paper titled A statistical approach to language translation. This marks a shift from rule-based to probabilistic methods of machine translation. It reflects a broader transition towards "machine learning" based on statistical analysis of known examples rather than a deep understanding of the task. IBM's project Candide, which successfully translates between English and French, relies on a massive dataset of 2.2 million pairs of sentences, primarily from the bilingual proceedings of the Canadian parliament.[10]
1988 German Research Centre for Artificial Intelligence Germany
1989 Marvin Minsky and Seymour Papert publish an expanded edition of their 1969 book Perceptrons. In a prologue added to the 1988 edition, they point out that progress in the field of artificial intelligence has been slow due to researchers repeating past mistakes, often because they were unaware of the field's history.[10]
1989 Yann LeCun and a team of researchers at AT&T Bell Labs achieve success by applying a backpropagation algorithm to a multi-layer neural network. This network is used to recognize handwritten ZIP codes. Despite hardware limitations at the time, the training of the network takes approximately three days, marking a significant improvement compared to earlier efforts.[10] United States
1989 Literature Journal of Experimental and Theoretical Artificial Intelligence
1989 (November 9) Literature The Emperor's New Mind: Concerning Computers, Minds and The Laws of Physics
1989 Dean Pomerleau at Carnegie Mellon University develops ALVINN (An Autonomous Land Vehicle in a Neural Network). This system would evolve into the technology that enables a car to be driven across the United States under computer control, with human intervention only required for about 50 of the 2850 miles of the journey.[8]
1990 Rodney Brooks publishes Elephants Don't Play Chess, advocating a novel approach to AI. His idea is to construct intelligent systems, particularly robots, by starting from the basics and allowing them to learn through continuous physical interaction with their environment. This approach emphasizes the importance of the real world as a model for intelligence and highlighted the need for effective and frequent sensory perception.[10]
1991 European Neural Network Society[44][45]
1991 American philanthropist Hugh Loebner starts the annual Loebner Prize competition, promising a $100,000 payout to the first computer to pass the Turing test and awarding $2,000 each year to the best effort. However, no AI program would come close to passing an undiluted Turing test.[16]
1992 Literature International Journal on Artificial Intelligence Tools[46]
1993 Journal of Artificial Intelligence Research[47]
1993 Vernor Vinge publishes "The Coming Technological Singularity," in which he forecasts that within thirty years, humanity would possess the technology to generate superhuman intelligence. He further anticipates that shortly after achieving this, the era of human dominance would come to an end.[10]
1994 (September) Conference The first Artificial Evolution Conference is held in Toulouse, France. It is the first international conference dedicated to the field of artificial evolution.[48] The conference is organized by the French Artificial Life Society (Société Française d'Évolution Artificielle) and the European Neural Networks Society (ESANN). The main topics of the conference were genetic algorithms, evolutionary programming, and evolutionary strategies. France
1995 Richard Wallace develops the chatbot A.L.I.C.E (Artificial Linguistic Internet Computer Entity), inspired by Joseph Weizenbaum's ELIZA program. A.L.I.C.E incorporates natural language sample data collected on an unprecedented scale, made possible by the advent of the World Wide Web.[10]
1995 A computer program called Chinook defeates the world checkers champion, Marion Tinsley, in a series of matches. Chinook uses a brute-force approach to checkers, evaluating all possible moves and selecting the best one. This approach is very computationally expensive, but becomes ultimately successful.[49]
1995 AltaVista becomes the first search engine to incorporate natural language processing into its functionality, enabling users to search for information using more human-like language and queries.[20]
1996 The EQP theorem prover at Argonne National Labs successfully proves the Robbins Conjecture in mathematics.[8]
1997 Sepp Hochreiter and Jürgen Schmidhuber propose Long Short-Term Memory (LSTM), a type of recurrent neural network that is widely used today in applications such as handwriting recognition and speech recognition.[10]
1997 IBM's Deep Blue chess computer defeates the reigning world chess champion, Garry Kasparov, in a six-game match. This is a major milestone in the field of artificial intelligence, as it shows that machines can now compete with humans at the highest level of chess.[49][4][10]
1997 Speech recognition software developed by Dragon Systems is implemented on Windows, marking significant progress in the field of spoken language interpretation.[4]
1998 Furby, the first domestic or pet robot, is created by Dave Hampton and Caleb Chung.[4]
1998 Literature Autonomous Agents and Multi-Agent Systems[50]
1998 Yann LeCun, Yoshua Bengio, and other researchers published papers on the application of neural networks to handwriting recognition and the optimization of backpropagation. These contributions were instrumental in advancing the field of neural network-based handwriting recognition.[4]
1998 Amazon introduces "collaborative filtering" to provide recommendations for millions of customers, a significant advancement in personalized recommendation systems.[51]
1998 Tiger Electronics releases Furby, marking the first successful introduction of AI technology into a domestic environment.[14]
Late 1990s Web crawlers and other AI-based information extraction programs become essential tools for the widespread use of the World Wide Web.[14]
1990s MIT's AI Lab demonstrates an Intelligent Room and Emotional Agents, showcasing advancements in intelligent environments and emotionally responsive agents. This period also marks the initiation of work on the Oxygen Architecture, which aims to connect mobile and stationary computers in an adaptive network, contributing to the development of pervasive computing.[7]
2000 MIT researcher Cynthia Breazeal develops Kismet, a robot capable of recognizing and simulating emotions, marking a significant advancement in emotional AI and human-robot interaction.[4][7]
2000 Honda's ASIMO robot, a humanoid robot endowed with artificial intelligence, achieves the capability to walk at a human-like speed and serve trays to customers in a restaurant setting, demonstrating significant progress in robotics and AI technology.[4]
2000 Conference Mexican International Conference on Artificial Intelligence[52] Mexico
2001 Artificial General Intelligence Research Institute[53] United States
2002 AI technology enters people's homes with the introduction of Roomba, an autonomous robotic vacuum cleaner. This marked a significant development in the application of AI to consumer products for everyday use.[1]
2002 Conference RuleML Symposium[54]
2003 Geoffrey Hinton, Yoshua Bengio, and Yann LeCun initiate a research program aimed at advancing neural networks. Experiments conducted in collaboration with Microsoft, Google, and IBM, with support from the Toronto laboratory led by Hinton, demonstrate significant improvements in speech recognition, effectively reducing error rates by half. Similar progress is achieved by Hinton's team in the field of image recognition. This marks a significant milestone in the development of neural network-based AI technologies.[3]
2003 MIT Computer Science and Artificial Intelligence Laboratory[55] United States
2004 The first DARPA Grand Challenge takes place, featuring a prize competition for autonomous vehicles. Unfortunately, none of the autonomous vehicles are able to complete the challenging 150-mile route in the Mojave Desert.[4]
2004 Conference International Conference on Computational Intelligence Methods for Bioinformatics and Biostatistics[56] Italy
2006 Oren Etzioni, Michele Banko, and Michael Cafarella introduce the term "machine reading," defining it as the autonomous understanding of text without the need for human supervision.[4]
2006 Geoffrey Hinton publishes a paper titled "Learning Multiple Layers of Representation," which summarizes ideas related to multilayer neural networks with top-down connections. This work represents a new approach to deep learning, focusing on training networks to generate sensory data rather than just classifying it.[4]
2006 AI begins to make its presence felt in the business world, with companies like Facebook, Twitter, and Netflix starting to utilize AI technologies for various purposes.[1]
2006 The first AI doctor-conducted unassisted robotic surgery is on a 34-year-old male to correct heart arrythmia. The results are rated as better than an above-average human surgeon. The machine has a database of 10,000 similar operations, and so, in the words of its designers, is "more than qualified to operate on any patient".[57][58]
2006 Conference AI@50, also known as the Dartmouth Artificial Intelligence Conference: The Next Fifty Years, takes place, marking the 50th anniversary of the Dartmouth workshop that initiated AI history. It features five of the original ten attendees, including Marvin Minsky and John McCarthy. The conference, sponsored by Dartmouth College, General Electric, and the Frederick Whittemore Foundation, receives a $200,000 grant from DARPA. Its goals include assessing AI's progress, identifying future challenges, and relating these to other fields. Conference topics range from emotion in machines to machine learning, vision, reasoning, and ethics.[59] United States
2007 Fei Fei Li and her team at Princeton University initiate the creation of ImageNet, a substantial database of annotated images intended to support research in visual object recognition software.[4] United States
2008 Eliezer Yudkowsky calls for the creation of “friendly AI” to mitigate existential risk from advanced artificial intelligence. Yudkowsky explains: "The AI does not hate you, nor does it love you, but you are made out of atoms which it can use for something else."[60] United States
2008 Conference Conference on Artificial General Intelligence[61]
2009 Rajat Raina, Anand Madhavan, and Andrew Ng publish Large-scale Deep Unsupervised Learning using Graphics Processors. They assert that modern graphics processors had significantly greater computational power compared to multicore CPUs and had the potential to revolutionize the use of deep unsupervised learning methods.[4]
2009 Google initiates the development of a driverless car project, which is kept confidential. By 2014, it would achieve a significant milestone by becoming the first to pass a self-driving test in the U.S. state of Nevada.[4]
2009 Computer scientists at Northwestern University's Intelligent Information Laboratory develop Stats Monkey, a program capable of autonomously generating sports news articles without any human involvement.[4]
2010 The ImageNet Large Scale Visual Recognition Challenge (ILSVRC) is launched as an annual competition focused on AI object recognition.[4]
2010 DeepMind[62]
2011 A convolutional neural network (CNN) achieved a remarkable victory in the German Traffic Sign Recognition competition by achieving an accuracy rate of 99.46%, surpassing the performance of human participants who scored 99.22%.[4]
2011 IBM's question-answering system, Watson, achieves a significant milestone by winning the quiz show "Jeopardy!" This victory occurs when Watson defeates the reigning champions, Brad Rutter and Ken Jennings.[5][4]
2011 A talking computer chatbot named Eugene Goostman gains attention for successfully deceiving judges into believing it was a genuine human during a Turing test.[5]
2011 Researchers at the IDSIA in Switzerland report a 0.27% error rate in handwriting recognition using convolutional neural networks in 2011. This is a significant improvement over the 0.35%-0.40% error rate in previous years.[4]
2011 A study published in the journal Nature Medicine shows that a machine learning algorithm called BioMind is able to outperform radiologists in interpreting breast cancer scans. The algorithm is trained on a dataset of over 100,000 scans, and is able to identify cancer with a 99% accuracy rate, compared to 96% for radiologists.[49]
2011 Apple's Siri is first released as part of the iPhone 4S. It is a major breakthrough in the field of artificial intelligence, as it is the first voice-activated personal assistant that is widely available.[23]
2012 (June) Jeff Dean and Andrew Ng conduct an experiment where they expose a massive neural network to 10 million unlabeled images randomly sourced from YouTube videos. Surprisingly, during this experiment, one of the artificial neurons within the network learns to respond strongly to images of cats, leading to an unexpected and amusing result.[4]
2012 (July 13) Literature The Machine Question: Critical Perspectives on AI, Robots, and Ethics
2012 Researchers at the University of Toronto develop a convolutional neural network that achieves a remarkable error rate of only 16% in the ImageNet Large Scale Visual Recognition Challenge. This marks a significant improvement compared to the previous year's best entry, which has an error rate of 25%.[10] Canada
2012 The secutiry market is flooded by computer vision start-ups.[63]
2013 Boston Dynamics unveils Atlas, an advanced humanoid robot designed for various search-and-rescue tasks. The robot is developed for the DARPA Robotics Challenge, a competition to develop robots that can perform tasks in disaster zones.[33][64] United States
2013 Automated Insights publish 300 million pieces of content, which Mashable reports is greater than the output of all major media companies combined. In 2014, the company's software would generate one billion stories. In 2016, Automated Insights would publish over 1.5 billion pieces of content.[51]
2014 (January) Google acquires DeepMind for a reported £400 million (about $650 million at the time).[65]
2014 Google starts developing a self-driving car in secret. The project is called "Project Chauffeur". In 2014, the project would be renamed to "Waymo".[4]
2014 Allen Institute for AI[66][67] United States
2014 A research team from the Chinese University of Hong Kong (CUHK) develops a facial recognition system that is able to achieve a human-level accuracy of 97.53%. This system is able to identify faces from a variety of angles and lighting conditions, and it is even able to identify faces that has been obscured by sunglasses or a mask.[49] China (Hong Kong)
2014 Microsoft introduces Cortana, a virtual assistant software. Cortana is first released for Windows Phone 8.1, and it is later released for Windows 10, Windows 10 Mobile, Xbox One, and Android.[23]
2014 Future of Life Institute[68] United States
2014 Squirrel AI[69][70] China
2014 Kiev Laboratory for Artificial Intelligence[71] Ukraine
2014 Ian Goodfellow introduces Generative Adversarial Networks (GAN), a revolutionary concept in artificial intelligence that involves two neural networks, a generator, and a discriminator, engaged in a competitive learning process to generate realistic data.[72]
2014 The rise of programmatic ad buying popularizes artificial intelligence-based ad purchasing. This innovation eliminates the need for time-consuming manual tasks such as market research, budgeting, insertion orders, and complex analytics tracking, making the ad buying process more efficient and cost-effective.[51]
2015 Amazon introduces the Alexa service. The first device to use Alexa is the Amazon Echo, a smart speaker that is released in June. Alexa is a cloud-based voice service that can be used to control smart home devices, play music, get news and weather updates, set alarms, and more. It would since become one of the most popular voice assistants in the world, with over 300 million active users.[23]
2015 (march) The algorithm for diffusion that would later serve as the foundation for text-to-image tools is initially introduced by researchers from Stanford and Berkeley.
2015 The Chinese Congress on Artificial Intelligence 2015 takes place in Beijing, giving the direction of AI-related industries in China.[63] China
2015 Open Letter on Artificial Intelligence[73]
2015 (September 22) The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World
2015 Google launches RankBrain, an advanced artificial intelligence algorithm. RankBrain would revolutionize search query interpretation by effectively understanding the user's search intent, resulting in more relevant search results.[51]
2016 (March) Google DeepMind's AlphaGo defeates Go champion Lee Sedol. This is a major milestone in the development of artificial intelligence, as Go is a much more complex game than chess.[4]
2016 (March) Microsoft releases the Tay chatbot, but quickly takes it offline after it becomes Holocaust denying.
2016 A team of researchers from Google AI and the University of Washington develops a machine learning model that can transcribe telephone calls with 97% accuracy. This is a significant improvement over previous methods, which has an accuracy of around 85%.[49]
2016 A team of researchers from the University of Oxford develops a machine learning model that can lipread with 94% accuracy. This is a significant improvement over previous methods, which has an accuracy of around 80%.[49]
2016 Center for Human-Compatible Artificial Intelligence[74] United States
2016 (February 16) Active Intelligence Pte Ltd[75] Singapore
2016 (September 28) Partnership on AI (full name Partnership on Artificial Intelligence to Benefit People and Society) is established. It is a non-profit partnership of academic, civil society, industry, and media organizations creating solutions so that AI advances positive outcomes for people and society.[76][77] Its founding members are Amazon, Facebook, Google, DeepMind, Microsoft, and IBM, with interim co-chairs Eric Horvitz of Microsoft Research and Mustafa Suleyman of DeepMind.[78][79] Apple would join the consortium as a founding member in January 2017.[80] By 2019, more than 100 partners from academia, civil society, industry, and nonprofits would be member organizations.[81]
2016 A real-time online tool called Swarm AI successfully predicts the winner of the Kentucky Derby horse race. This demonstrates the potential of collective intelligence and real-time collaboration among users to make accurate predictions.[33]
2017 OpenAI Five[82] United States
2017 DeepMind releases AI Safety Gridworlds, which evaluate AI algorithms on nine safety features, such as whether the algorithm wants to turn off its own kill switch. DeepMind confirms that existing algorithms perform poorly, which is "unsurprising" because the algorithms "are not designed to solve these problems"; solving such problems might require "potentially building a new generation of algorithms with safety considerations at their core".[83][84][85]
2017 Conference The Asilomar Conference on Beneficial AI isn held, focusing on discussing the potential risks and benefits associated with artificial intelligence (AI) and how to ensure the development of AI in a way that benefits humanity.[86]
2017 The first AI for Good Global summit takes place.[87]
2017 Organization AI Now Institute is founded. It is an American research institute studying the social implications of artificial intelligence.[88] United States
2017 The AI market, including both hardware and software, reaches a total value of $8 billion.[37]
2017 A convolutional neural network (CNN) achieves a remarkable victory in the German Traffic Sign Recognition competition by achieving an accuracy rate of 99.46%, surpassing the performance of human participants who scored 99.22%.[33]
2017 Google's DeepMind AI achieves the remarkable feat of teaching itself how to walk autonomously.[33]
2017 AI is included in the Chinese government report as a national strategy in China.[63]
2018 Artificial intelligence showcases its abilities in different ways. IBM's 'Project Debater' engaged in complex debates with human master debaters and performed impressively. Meanwhile, Google's 'Duplex' AI demonstrated its conversational skills by making a hairdressing appointment over the phone without the recipient realizing they were talking to a machine. These examples illustrated AI's capacity to tackle advanced tasks and engage in natural conversations.[23]
2018 A machine learning algorithm called BioMind is able to outperform radiologists in interpreting breast cancer scans. The algorithm is trained on a dataset of over 100,000 scans, and is able to identify cancer with a 99% accuracy rate, compared to 96% for radiologists.[49]
2018 European Laboratory for Learning and Intelligent Systems[89]
2018 (April 26) Innovation Center for Artificial Intelligence[90][91] Netherlands
2018 The artificial intelligence market in China amounts to 33.9 billion RMB.[63] China
2018 Astronomers harness the power of AI to identify and locate approximately 6,000 new craters on the moon's surface, enhancing our understanding of lunar geology.[33][92]
2018 Paul Rad, assistant director of the University of Texas-San Antonio Open Cloud Institute, and Nicole Beebe, director of the university's Cyber Center for Security and Analytics, introduce a novel cloud-based learning platform for AI. This platform aims to teach machines to learn in a manner similar to human learning processes.[33][93]
2018 Google showcases Duplex AI, a digital assistant capable of making appointments via telephone calls with live humans. Duplex utilizes natural language understanding, deep learning, and text-to-speech technologies to grasp conversational context and nuances, achieving a level of sophistication unmatched by other digital assistants.[33]
2018 AI ushers in the first year of commercial applications in China. There are more than 1,000 AI-related companies in the country by the time.[63] China
2018 The AI Now Report finds harmful inaccuracies in AI-driven technology, plus an alarming lack of accountability and, in some cases, systems built on racial discrimination or used for human rights violations.[94]
2019 Center for Security and Emerging Technology[95][96] United States
2019 Google AI Centre in Ghana[97][98] Ghana
2019 A team of five AI bots developed by OpenAI called OpenAI Five defeates a team of professional Dota 2 players in a best-of-three match. This is a significant achievement, as Dota 2 is a complex multiplayer game that requires a high degree of teamwork and strategy.[49]
2019 AI Artathon[99][100] Saudi Arabia
2020 An AI called Agent57 developed by DeepMind is able to beat humans at all 57 Atari 2600 games. This is a significant achievement, as the Atari 2600 is a classic console with a wide range of challenging games.[49]
2020 (June) OpenAI reveals GPT-3, but releases it only to a small pool of users.

Meta information on the timeline

How the timeline was built

The initial version of the timeline was written by User:Sebastian.

Funding information for this timeline is available.

Feedback and comments

Feedback for the timeline can be provided at the following places:

  • FIXME

What the timeline is still missing

Timeline update strategy

See also

External links

References

  1. 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 "History of Artificial Intelligence". javatpoint.com. Retrieved 7 February 2020. 
  2. 2.0 2.1 2.2 2.3 2.4 2.5 2.6 "A Brief History of Artificial Intelligence". dataversity.net. Retrieved 7 February 2020. 
  3. 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 Cite error: Invalid <ref> tag; no text was provided for refs named coe.intf
  4. 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 4.18 4.19 4.20 4.21 4.22 4.23 4.24 4.25 "The History of Artificial Intelligence". harvard.edu. Retrieved 7 February 2020. 
  5. 5.0 5.1 5.2 5.3 5.4 5.5 "A Brief History of Artificial Intelligence". livescience.com. Retrieved 7 February 2020. 
  6. 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 "The History of Artificial Intelligence" (PDF). washington.edu. Retrieved 7 February 2020. 
  7. 7.0 7.1 7.2 "Tema 1 Brief History of Artificial Intelligence". ocw.uc3m.es. Retrieved 21 March 2020. 
  8. 8.00 8.01 8.02 8.03 8.04 8.05 8.06 8.07 8.08 8.09 8.10 8.11 8.12 8.13 8.14 8.15 8.16 8.17 8.18 8.19 8.20 8.21 8.22 8.23 8.24 8.25 8.26 8.27 8.28 8.29 8.30 8.31 8.32 8.33 8.34 8.35 8.36 8.37 8.38 8.39 8.40 8.41 8.42 8.43 8.44 8.45 8.46 8.47 8.48 8.49 8.50 8.51 8.52 8.53 8.54 8.55 8.56 8.57 8.58 8.59 8.60 8.61 "A Brief History of AI". aitopics.org. Retrieved 20 March 2020. 
  9. 9.00 9.01 9.02 9.03 9.04 9.05 9.06 9.07 9.08 9.09 9.10 9.11 9.12 9.13 Mijwil, Maad M. "History of Artificial Intelligence". Retrieved 9 March 2020. 
  10. 10.00 10.01 10.02 10.03 10.04 10.05 10.06 10.07 10.08 10.09 10.10 10.11 10.12 10.13 10.14 10.15 10.16 10.17 10.18 10.19 10.20 10.21 10.22 10.23 10.24 10.25 10.26 10.27 10.28 10.29 10.30 10.31 10.32 10.33 10.34 10.35 10.36 10.37 10.38 10.39 10.40 10.41 10.42 10.43 10.44 10.45 10.46 10.47 10.48 10.49 10.50 10.51 10.52 10.53 10.54 10.55 10.56 "A Very Short History Of Artificial Intelligence (AI)". forbes.com. Retrieved 7 February 2020. 
  11. Mehta, Dhaval; Ranadive, Dr Amol (31 January 2021). What Gamers Want: A Framework to Predict Gaming Habits. OrangeBooks Publication. 
  12. 12.0 12.1 Bloch, Laurent. "Informatics in the light of some Leibniz's works" (PDF). laurentbloch.net. Retrieved 9 March 2022. 
  13. Kumar, Ajitesh (17 September 2021). "12 Bayesian Machine Learning Applications Examples". Data Analytics. Retrieved 7 March 2022. 
  14. 14.0 14.1 14.2 "The History Of Artificial Intelligence". sutori.com. Retrieved 20 March 2020. 
  15. "Artificial Intelligence". people.idsia.ch. Retrieved 21 March 2020. 
  16. 16.00 16.01 16.02 16.03 16.04 16.05 16.06 16.07 16.08 16.09 16.10 "Artificial intelligence". britannica.com. Retrieved 21 March 2020. 
  17. "A BRIEF HISTORY OF ARTIFICIAL INTELLIGENCE". atariarchives.org. Retrieved 21 March 2020. 
  18. "History of Artificial Intelligence". researchgate.net. Retrieved 9 March 2020. 
  19. 19.0 19.1 19.2 "1.2 A Brief History of Artificial Intelligence". artint.info. Retrieved 21 March 2020. 
  20. 20.0 20.1 "A SHORT HISTORY OF ARTIFICIAL INTELLIGENCE: MAKING MYTHOLOGY A REALITY". omnius.com. Retrieved 20 March 2020. 
  21. Engineers: From the Great Pyramids to the Pioneers of Space Travel. Penguin. 16 April 2012. ISBN 978-1-4654-0682-8. 
  22. "7 phases of the history of Artificial intelligence". historyextra.com. Retrieved 21 March 2020. 
  23. 23.0 23.1 23.2 23.3 23.4 23.5 23.6 "The history of artificial intelligence". bosch.com. Retrieved 7 February 2020. 
  24. "AIC Timeline". ai.sri.com. Retrieved 6 March 2020. 
  25. "Semantic Network - an overview | ScienceDirect Topics". www.sciencedirect.com. Retrieved 5 March 2022. 
  26. "Notes on Semantic Nets and Frames" (PDF). eecs.qmul.ac.uk. Retrieved 5 March 2022. 
  27. "Artificial Intelligence Journal Division of IJCAI". ijcai.org. Retrieved 6 March 2020. 
  28. Harris, Randy Allen (31 December 2004). Voice Interaction Design: Crafting the New Conversational Speech Systems. Elsevier. ISBN 978-0-08-047480-9. 
  29. Shapiro, Stuart C. (1 January 1982). "Generalized augmented transition network grammars for generation from semantic networks". Computational Linguistics. 8 (1): 12–25. ISSN 0891-2017. doi:10.5555/972923.972925. 
  30. Dreyfus, Hubert L. (30 October 1992). "What Computers Still Can't Do: A Critique of Artificial Reason". mitpress.mit.edu. MIT Press. Retrieved 21 March 2022. 
  31. Eng, Richard Kenneth (23 July 2022). "Celebrating 50 Years of Smalltalk". Medium. Retrieved 9 September 2023. 
  32. "ECAI 2010". iospress.nl. Retrieved 6 March 2020. 
  33. 33.0 33.1 33.2 33.3 33.4 33.5 33.6 33.7 33.8 "The History of Artificial Intelligence". futureoftech.org. Retrieved 9 March 2020. 
  34. "ILabs". semanticscholar.org. Retrieved 6 March 2020. 
  35. "The Association for the Advancement of Artificial Intelligence (AAAI)". www.omicsonline.org. Retrieved 21 March 2022. 
  36. "History of Artificial Intelligence – AI of the past, present and the future!". data-flair.training. Retrieved 4 March 2020. 
  37. 37.0 37.1 "A Short History of Artificial Intelligence". dev.to. Retrieved 9 March 2020. 
  38. Lamb, John (August 1985). Making Friends with Intelligence. The New Scientist. Retrieved 10 December 2013. 
  39. "Column 468: The Turing Institute". UK Parliament. Retrieved 2 March 2022. 
  40. Aydin, Berkay; Angryk, Rafal A. (15 October 2018). Spatiotemporal Frequent Pattern Mining from Evolving Region Trajectories. Springer. ISBN 978-3-319-99873-2. 
  41. Liang-Jie, Zhang; Yishuang, Ning (19 October 2018). Innovative Solutions and Applications of Web Services Technology. IGI Global. ISBN 978-1-5225-7269-5. 
  42. "Qualitative Spatio-Temporal Reasoning with RCC-8 and Allen's Interval Calculus: Computational Complexity" (PDF). gki.informatik.uni-freiburg.de. Retrieved 12 March 2022. 
  43. "Centre for Artificial Intelligence and Robotics (CAIR)". epicos.com. Retrieved 6 March 2020. 
  44. Taylor, J.G. The Promise of Neural Networks. 
  45. Artificial Neural Networks and Machine Learning – ICANN 2017: 26th International Conference on Artificial Neural Networks, Alghero, Italy, September 11-14, 2017, Proceedings, Part 1 (Alessandra Lintas, Stefano Rovetta, Paul F.M.J. Verschure, Alessandro E.P. Villa ed.). 
  46. "International Journal on Artificial Intelligence Tools". letpub.com. Retrieved 6 March 2020. 
  47. "Journal of Artificial Intelligence Research". jair.org. Retrieved 6 March 2020. 
  48. "Artificial Evolution 2019 (EA-2019)". iscpif.fr. Retrieved 6 March 2020. 
  49. 49.0 49.1 49.2 49.3 49.4 49.5 49.6 49.7 49.8 Leigh, Andrew (9 November 2021). What's the Worst That Could Happen?: Existential Risk and Extreme Politics. MIT Press. ISBN 978-0-262-36661-8. 
  50. "Autonomous Agents and Multi-Agent Systems". springer.com. Retrieved 6 March 2020. 
  51. 51.0 51.1 51.2 51.3 "A brief history of artificial intelligence in advertising". econsultancy.com. Retrieved 20 March 2020. 
  52. "MICAI 2000: Advances in Artificial Intelligence". springer.com. Retrieved 6 March 2020. 
  53. "Artificial General Intelligence Research Institute". morebooks.de. Retrieved 6 March 2020. 
  54. Bikakis, Antonis; Fodor, Paul; Roman, Dumitru. Rules on the Web: From Theory to Applications: 8th International Symposium, RuleML 2014, Co-located with the 21st European Conference on Artificial Intelligence, ECAI 2014, Prague, Czech Republic, August 18-20, 2014, Proceedings. 
  55. "Mission & History". csail.mit.edu. Retrieved 6 March 2020. 
  56. "INTERNATIONAL MEETING ON COMPUTATIONAL INTELLIGENCE METHODS FOR BIOINFORMATICS AND BIOSTATISTICS". person.dibris.unige.it. Retrieved 6 March 2020. 
  57. "Autonomous Robotic Surgeon performs surgery on first live human". Engadget. 19 May 2006. 
  58. "Robot surgeon carries out 9-hour operation by itself". Phys.Org. 
  59. "Dartmouth Artificial Intelligence Conference". dartmouth.edu. Retrieved 6 March 2020. 
  60. Eliezer Yudkowsky (2008) in Artificial Intelligence as a Positive and Negative Factor in Global Risk
  61. "Artificial General Intelligence 2008". iospress.nl. Retrieved 6 March 2020. 
  62. "Expanding our knowledge, finding new answers". deepmind.com. Retrieved 6 March 2020. 
  63. 63.0 63.1 63.2 63.3 63.4 "The history of Artificial Intelligence (AI) in China". daxueconsulting.com. Retrieved 21 March 2020. 
  64. "Atlas". bostondynamics.com. Retrieved 9 March 2020. 
  65. "A Brief History of Artificial Intelligence". kdnuggets.com. Retrieved 9 March 2020. 
  66. "Allen Institute for AI". glassdoor.com.ar. Retrieved 6 March 2020. 
  67. "Allen Institute for AI (AI2)". linkedin.com. Retrieved 6 March 2020. 
  68. "Future of Life Institute". linkedin.com. Retrieved 6 March 2020. 
  69. "Adaptive Learning Startup Squirrel AI Raises CN¥1B". medium.com. Retrieved 6 March 2020. 
  70. "Squirrel AI Learning". crunchbase.com. Retrieved 6 March 2020. 
  71. "Kiev Laboratory for Artificial Intelligence". semanticscholar.org. Retrieved 6 March 2020. 
  72. "History of Artificial Intelligence". qbi.uq.edu.au. Retrieved 9 March 2020. 
  73. "Elon Musk, Stephen Hawking warn of artificial intelligence dangers". mashable.com. Retrieved 6 March 2020. 
  74. "UC Berkeley launches Center for Human-Compatible Artificial Intelligence". news.berkeley.edu. Retrieved 6 March 2020. 
  75. "ACTIVE INTELLIGENCE PTE. LTD.". sgpbusiness.com. Retrieved 6 March 2020. 
  76. "Exploring The Partnership on AI". medium.com. Retrieved 6 March 2020. 
  77. "About". Partnership on AI. Retrieved 3 March 2022. 
  78. "About". Partnership on AI. Retrieved 3 March 2022. 
  79. "'Partnership on AI' formed by Google, Facebook, Amazon, IBM and Microsoft". the Guardian. 28 September 2016. Retrieved 3 March 2022. 
  80. "Partnership on AI Update". Partnership on AI. Retrieved 3 March 2022. 
  81. "New Partners To Bolster Perspective For Responsible AI". Partnership on AI. 24 September 2019. Retrieved 3 March 2022. 
  82. "OpenAI Five". openai.com. Retrieved 6 March 2020. 
  83. "DeepMind Has Simple Tests That Might Prevent Elon Musk's AI Apocalypse". Bloomberg.com. 11 December 2017. Retrieved 5 March 2020. 
  84. "Alphabet's DeepMind Is Using Games to Discover If Artificial Intelligence Can Break Free and Kill Us All". Fortune. Retrieved 5 March 2020. 
  85. "Specifying AI safety problems in simple environments | DeepMind". DeepMind. Retrieved 5 March 2020. 
  86. "Video: Superintelligence Panel at Beneficial AI 2017 (FLI)". medium.com. Retrieved 6 March 2020. 
  87. "AI for Good Global Summit 2017". ITU. Retrieved 10 March 2023. 
  88. "NYU Law and NYU's AI Now Institute analyze the ways emerging technology imposes upon civil liberties". law.nyu.edu. Retrieved 6 March 2020. 
  89. "European Laboratory for Learning and Intelligent Systems (ELLIS) launched with Informatics researchers on board". ed.ac.uk. Retrieved 9 March 2020. 
  90. "Innovation Center for Artificial Intelligence officially launched". uva.nl. Retrieved 6 March 2020. 
  91. "Ahold Delhaize Helps Launch AI Innovation Center". consumergoods.com. Retrieved 6 March 2020. 
  92. "New technique uses AI to locate and count craters on the moon". phys.org. Retrieved 9 March 2020. 
  93. "UTSA researchers want to teach computers to learn like humans". utsa.edu. Retrieved 9 March 2020. 
  94. "Rise of the Machines: The History of Artificial Intelligence". looklisten.com. Retrieved 21 March 2020. 
  95. "Center for Security and Emerging Technology". cset.georgetown.edu. Retrieved 6 March 2020. 
  96. "Center for Security and Emerging Technology". linkedin.com. Retrieved 6 March 2020. 
  97. "Google takes on 'Africa's challenges' with first AI centre in Ghana". thestar.com.my. Retrieved 6 March 2020. 
  98. "How Google is driving artificial intelligence for Africa by Africans". techpoint.africa. Retrieved 6 March 2020. 
  99. "About the Global AI Summit". theglobalaisummit.com. Retrieved 6 March 2020. 
  100. "Riyadh to host AI art competition". arabnews.jp. Retrieved 6 March 2020.