The Artificial Intelligence Chronicle

384 BC–322 BC Aristotle described the syllogism, a method of formal, mechanical  thought and theory of knowledge, the first formal deductive reasoning system. in The  Organon

1st century Heron of Alexandria created mechanical men and other automatons. 260 Porphyry of Tyros wrote Isagogê which categorized knowledge and logic. ~800 Geber developed the Arabic alchemical theory of Takwin, the artificial creation of life  in the laboratory, up to and including human life.  

1206 Al-Jazari created a programmable orchestra of mechanical human beings. 

Talking heads were said to have been created, Roger Bacon and Albert the  Great reputedly among the owners.  

In 1206 A.D., Al-Jazari, an Arab inventor, designed what is believed to be the first  programmable humanoid robot, a boat carrying four mechanical musicians  powered by water flow. 

1275 Ramon Llull, Spanish theologian, invents the Ars Magna, a tool for combining concepts  mechanically, based on an Arabic astrological tool, the Zairja. The method would be developed further  by Gottfried Leibniz in the 17th century. 

1308 Catalan poet and theologian Ramon Llull publishes Ars  generalis ultima (The Ultimate General Art), further perfecting his method of  using paper-based mechanical means to create new knowledge from  combinations of concepts. 

15th century 

  • Invention of printing using moveable type. Gutenberg Bible printed (1456).

15th-16th century

  • Clocks, the first modern measuring machines, were first produced using lathes.

~1500 Paracelsus claimed to have created an artificial man out of magnetism, sperm and  alchemy.

16th century 

  • Clockmakers extended their craft to creating mechanical animals and other novelties. For example, see DaVinci’s walking lion (1515).

~1580 Rabbi Judah Loew ben Bezalel of Prague is said to have invented the Golem, a  clay man brought to life.

Early 1700s: Depictions of all-knowing machines akin to computers were more  widely discussed in popular literature. Jonathan Swift’s novel “Gulliver’s Travels”  mentioned a device called the engine, which is one of the earliest references to  modern-day technology, specifically a computer. This device’s intended purpose  was to improve knowledge and mechanical operations to a point where even the  least talented person would seem to be skilled – all with the assistance and  knowledge of a non-human mind (mimicking artificial intelligence.) 

  • Early in the century, Descartes proposed that bodies of animals are nothing more than complex machines. Many other 17th century thinkers offered variations and elaborations of Cartesian mechanism. 

Early 17th century René Descartes proposed that bodies of animals are nothing  more than complex machines (but that mental phenomena are of a different “substance”).

1620 Sir Francis Bacon developed empirical theory of knowledge and introduced  inductive logic in his work The New Organon, a play on Aristotle‘s title The Organon.

1623 Wilhelm Schickard drew a calculating clock on a letter to Kepler. This will be the  first of five unsuccessful attempts at designing a direct entry calculating clock in the 17th century  (including the designs of Tito Burattini, Samuel Morland and René Grillet).

1641 Thomas Hobbes published Leviathan and presented a mechanical, combinatorial  theory of cognition. He wrote “…for reason is nothing but reckoning”.

1642 Blaise Pascal invented the mechanical calculator, the first digital calculating  machine.

  • Thomas Hobbes published The Leviathan (1651), containing a mechanistic and combinatorial theory of thinking. 
  • Arithmetical machines devised by Sir Samuel Morland between 1662 and 1666 

1666 Mathematician and philosopher Gottfried Leibniz  

publishes Dissertatio de arte combinatoria (On the Combinatorial Art), following  Ramon Llull in proposing an alphabet of human thought and arguing that all  ideas are nothing but combinations of a relatively small number of simple  concepts. 

1672 Gottfried Leibniz improved the earlier machines, making the Stepped Reckoner to  do multiplication and division. He also invented the binary numeral system and envisioned a universal  calculus of reasoning (alphabet of human thought) by which arguments could be decided  mechanically. Leibniz worked on assigning a specific number to each and every object in the world, as a  prelude to an algebraic solution to all possible problems.


18th century 

1726 Jonathan Swift publishes Gulliver’s Travels, which  includes a description of the Engine, a machine on the island of Laputa (and a  parody of Llull’s ideas): “a Project for improving speculative Knowledge by  practical and mechanical Operations.” By using this “Contrivance,” “the most  ignorant Person at a reasonable Charge, and with a little bodily Labour, may write  Books in Philosophy, Poetry, Politicks, Law, Mathematicks, and Theology, with  the least Assistance from Genius or study.” The machine is a parody of Ars Magna, one of  the inspirations of Gottfried Leibniz‘ mechanism. (3) 

1750 Julien Offray de La Mettrie published L’Homme Machine, which argued that human thought is  strictly mechanical. (1) 

1763 Thomas Bayes develops a framework for reasoning about  the probability of events. Bayesian inference will become a leading approach in  machine learning. (1) 

1769 Wolfgang von Kempelen built and toured with his chess-playing automaton, The Turk.[24] The  Turk was later shown to be a hoax, involving a human chess player. (3) 

  • The 18th century saw a profusion of mechanical toys, including the celebrated mechanical duck of Vaucanson and von Kempelen’s phony mechanical chess player, The Turk (1769). Edgar Allen Poe wrote (in the Southern  Literary Messenger, April 1836) that the Turk could not be a machine because, if it  were, it would not lose. 

19th century 

  • Joseph-Marie Jacquard invented the Jacquard loom, the first programmable machine, with instructions on punched cards (1801). (2)
  • Luddites (by Marjie Bloy, PhD. Victorian Web) (led by Ned Ludd) destroyed machinery in England (1811-1816). See also What the Luddites Really Fought Against. By Richard Conniff, Smithsonian magazine (March 2011). (3) 
  • Mary Shelley published the story of Frankenstein’s monster (1818). The book Frankenstein, or the Modern Prometheus available from Project Gutenberg. (1)
  • Charles Babbage & Ada Byron (Lady Lovelace) designed a programmable mechanical calculating machines, the Analytical Engine (1832). A working model was built in 2002; a short video shows it working. 

1837 The mathematician Bernard Bolzano made the first modern attempt to formalize semantics. (1) 

1854 George Boole argues that logical reasoning could be  performed systematically in the same manner as solving a system of equations. George Boole developed a binary algebra representing (some) “laws of thought,”  published in The Laws of Thought (1854). (1) 

1863 Samuel Butler suggested that Darwinian evolution also applies to machines, and speculates that  they will one day become conscious and eventually supplant humanity. (1) 

  • 1872: Author Samuel Butler’s novel “Erewhon” toyed with the idea that at  an indeterminate point in the future machines would have the potential to  possess consciousness. (1) 
  • Modern propositional logic developed by Gottlob Frege in his 1879 work Begriffsschrift and later clarified and expanded by Russell,Tarski, Godel, Church and others. (1) 
  • 1898 At an electrical exhibition in the recently completed  Madison Square Garden, Nikola Tesla makes a demonstration of the world’s first radio-controlled vessel. The boat was equipped with, as Tesla  described, “a borrowed mind.” (3) 

AI from 1900-1950 

Once the 1900s hit, the pace with which innovation in artificial intelligence grew  was significant. 

  • Bertrand Russell and Alfred North Whitehead published Principia Mathematica, which revolutionaized formal logic. Russell, Ludwig Wittgenstein, and Rudolf Carnap lead philosophy into logical analysis of knowledge. (1)
  • 1914 The Spanish engineer Leonardo Torres y Quevedo demonstrates the first chess-playing machine, capable of king and rook against king endgames without any human intervention, his chess  machine ‘Ajedrecista’, using electromagnets under the board to play the  endgame rook and king against the lone king, possibly the first computer  game (1912). (3) 

1921: Karel Čapek, a Czech playwright, released his science fiction play  “Rossum’s Universal Robots” (English translation). His play explored the concept  of factory-made artificial people who he called robots – the first known reference  to the word. From this point onward, people took the “robot” idea and  implemented it into their research, art, and discoveries. (2)


1925 Houdina Radio Control releases a radio-controlled  driverless car, travelling the streets of New York City. (3) 

1927: The sci-fi film Metropolis, directed by Fritz Lang, featured a robotic girl who  was physically indistinguishable from the human counterpart from which it took its  likeness. The artificially intelligent robot-girl then attacks the town, wreaking  havoc on a futuristic Berlin. It features a robot double of a peasant girl, Maria,  which unleashes chaos in Berlin of 2026—it was the first robot depicted on film,  inspiring the Art Deco look of C-3PO in Star Wars. This film holds significance  because it is the first on-screen depiction of a robot and thus lent inspiration to  other famous non-human characters. (2)


1929: Japanese biologist and professor Makoto Nishimura created Gakutensoku,  the first robot to be built in Japan. Gakutensoku translates to “learning from the  laws of nature,” implying the robot’s artificially intelligent mind could derive  knowledge from people and nature. Some of its features included moving its head  and hands as well as changing its facial expressions and move its head and  hands via an air pressure mechanism. (3) 

1920s and 1930s Ludwig Wittgenstein and Rudolf Carnap led philosophy into logical analysis  of knowledge. Alonzo Church developde Lambda Calculus to investigate computability using recursive  functional notation.(1) 

1931 Kurt Gödel showed that sufficiently powerful formal systems, if consistent, permit the formulation  of true theorems that are unprovable by any theorem-proving machine deriving all possible theorems  from the axioms. To do this he had to build a universal, integer-based programming language, which is  the reason why he is sometimes called the “father of theoretical computer science“. (1)


Alan Turing proposed the universal Turing machine (1936-37) 

1939: John Vincent Atanasoff (physicist and inventor), alongside his graduate  student assistant Clifford Berry, created the Atanasoff-Berry Computer (ABC) with 


a grant of $650 at Iowa State University. The ABC weighed over 700 pounds and  could solve up to 29 simultaneous linear equations. 

1939 Electro, a mechanical man, introduced by Westinghouse Electricat the  World’s Fair in New York (1939), along with Sparko, a mechanical dog. Built by  Westinghouse, the relay-based Elektro robot responds to the  rhythm of voice commands and delivers wisecracks pre-recorded  on 78 rpm records. It appeared at the World’s Fair, and it could  move its head and arms… and even “smoked” cigarettes.


1940 Edward Condon displays Nimatron, a digital computer that played Nim perfectly. (3) 1941 Konrad Zuse built the first working program-controlled computers. (3) 

1941 The Three Laws of Robotics: Isaac Asimov publishes the  science fiction short story Liar! in the May issue of Astounding  Science Fiction. In it, he introduced the Three Laws of Robotics: 

A robot may not injure a human being or, through inaction,  allow a human being to come to harm.


A robot must obey the orders given to it by human beings,  except where such orders would conflict with the First Law. 

A robot must protect its own existence as long as such protection does not conflict  with the First or Second Laws. 

This is thought to be the first known use of the term “robotics.” (1,2)


1943 Warren S. McCulloch and Walter Pitts publish “A Logical  Calculus of the Ideas Immanent in Nervous Activity” in the Bulletin of  Mathematical Biophysics. This influential paper, in which they discussed  networks of idealized and simplified artificial “neurons” and how they might  perform simple logical functions, will become the inspiration for computer-based  “neural networks” (and later “deep learning”) and their popular description as  “mimicking the brain.” 

Two scientists, Warren S. McCulloch and Walter H. Pitts, publish  the groundbreaking paper A Logical Calculus of the Ideas Immanent in  Nervous Activity. The paper quickly became a foundational work in  the study of artificial neural networks and has many applications in  artificial intelligence research. In it McCulloch and Pitts described a  simplified neural network architecture for intelligence, and while the  neurons they described were greatly simplified compared to  biological neurons, the model they proposed was enhanced and  improved upon by subsequent generations of researchers. (1)


Emil Post proves that production systems are a general computational mechanism  (1943). See Ch.2 of Rule Based Expert Systems for the uses of production systems in  AI. Post also did important work on completeness, inconsistency, and proof  theory. (1) 

1945 Game theory which would prove invaluable in the progress of AI was introduced with the 1944  paper, Theory of Games and Economic Behavior by mathematician John von  

Neumann and economist Oskar Morgenstern. (1) 

1945 Vannevar Bush published As We May Think (The Atlantic Monthly, July 1945) a prescient vision  of the future in which computers assist humans in many activities. (1)


  • Arturo Rosenblueth, Norbert Wiener & Julian Bigelow coin the term “cybernetics” in a 1943 paper. Wiener’s popular book by that name published in 1948. (1) 
  • George Polya published his best-selling book on thinking heuristically, How to Solve It in 1945. This book introduced the term ‘heuristic’ into modern thinking and has influenced many AI scientists. (1) 
  • Vannevar Bush published As We May Think (Atlantic Monthly, July 1945) a prescient vision of the future in which computers assist humans in many activities. (1) 
  • Grey Walter experimented with autonomous robots, turtles named Elsie and Elmer, at Bristol (1948-49) based on the premise that a small number of brain cells could give rise to complex behaviors. (3) 
  • A.M. Turing published “Computing Machinery and Intelligence” (1950). – Introduction of Turing Test as a way of operationalizing a test of intelligent behavior. See The Turing Institute for more on Turing. (3) 

1948 Cybernetics: Norbert Wiener publishes the book Cybernetics,  which has a major influence on research into artificial intelligence  and control systems. Wiener drew on his World War II experiments  with anti-aircraft systems that anticipated the course of enemy  planes by interpreting radar images. Wiener coined the term  “cybernetics” from the Greek word for “steersman.” (1)

1948 John von Neumann (quoted by E.T. Jaynes) in response to a comment at a lecture that it was  impossible for a machine to think: “You insist that there is something a machine cannot do. If you will tell  me precisely what it is that a machine cannot do, then I can always make a machine which will do just 


that!”. Von Neumann was presumably alluding to the Church-Turing thesis which states that any effective  procedure can be simulated by a (generalized) computer. (1) 

1949 Alan Turing quoted by The London Times on artificial  intelligence: On June 11, The London Times quotes the  mathematician Alan Turing. “I do not see why it (the machine)  should not enter any one of the fields normally covered by the  human intellect, and eventually compete on equal terms. I do not  think you even draw the line about sonnets, though the comparison  is perhaps a little bit unfair because a sonnet written by a machine  will be better appreciated by another machine.” (1)



1949: Computer scientist Edmund Berkeley’s book “Giant Brains: Or Machines  That Think” in which he writes: “Recently there have been a good deal of news  about strange giant machines that can handle information with vast speed and  skill…. He went on to compare machines to a human brain if it were made of  “hardware and wire instead of flesh and nerves,” describing machine ability to  that of the human mind, stating that “a machine, therefore, can think.” (1) 

1949 Donald Hebb publishes Organization of Behavior: A  Neuropsychological Theory in which he proposes a theory about learning based  on conjectures regarding neural networks and the ability of synapses to  strengthen or weaken over time. Claude Shannon published detailed analysis of  chess playing as search in “Programming a computer to play chess” (1950). (1) 

AI in the 1950s 

The 1950s proved to be a time when many advances in the field of artificial  intelligence came to fruition with an upswing in research-based findings in AI by  various computer scientists among others. The modern history of AI begins with  the development of stored-program electronic computers. For a short summary,  see Genius and Tragedy at Dawn of Computer Age By ALICE RAWSTHORN, NY Times  (March 25, 2012 ), a review of technology historian George Dyson’s book  “Turing’s Cathedral: The Origins of the Digital Universe.” 

1950 Grey Walter’s Elsie: A neurophysiologist, Walter built  wheeled automatons in order to experiment with goal-seeking  behavior. His best known robot, Elsie, used photoelectric cells to  seek moderate light while avoiding both strong light and darkness— which made it peculiarly attracted to women’s stockings. (2,3)



1950 Isaac Asimov published his three laws of robotics (1950). (1) 

1950 Issac Asimov’s, Robot: Isaac Asimov’s I, Robot is published.  Perhaps in reaction to earlier dangerous fictional robots, Asimov’s  creations must obey the “Three Laws of Robotics” (1941) to assure  they are no threat to humans or each other. The book consisted of  nine science fiction short stories. (1)


1950 Brain surgeon reflects on artificial intelligence: On June 9, at  Manchester University’s Lister Oration, British brain surgeon  Geoffrey Jefferson states, “Not until a machine can write a sonnet 


or compose a concerto because of thoughts and emotions felt, and  not by the chance fall of symbols, could we agree that machine  equals brain – that is, not only write it but know that it had written it.  No mechanism could feel (and not merely artificially signal, an easy  contrivance) pleasure at its successes, grief when its valves fuse, be  warmed by flattery, be made miserable by its mistakes, be charmed  by sex, be angry or miserable when it cannot get what it wants.” (1)


1950: Claude Shannon, “the father of information theory,” published  “Programming a Computer for Playing Chess,” which was the first article to  discuss the development of a chess-playing computer program. (1) 

1950: Alan Turing published “Computing Machinery and Intelligence,” which  proposed the idea of The Imitation Game – a question that considered if 


machines can think. This proposal later became The Turing Test, which  measured machine (artificial) intelligence. Turing’s development tested a  machine’s ability to think as a human would. The Turing Test became an  important component in the philosophy of artificial intelligence, which discusses  intelligence, consciousness, and ability in machines. (1,2) 

1951 The first working AI programs were written in 1951 to run on the Ferranti Mark 1 machine of  the University of Manchester: a checkers-playing program written by Christopher Strachey and a chess playing program written by Dietrich Prinz. (3) 

1951 Squee: the robot squirrel: Squee: The Robot Squirrel uses two  light sensors and two contact switches to hunt for ”nuts” (actually,  tennis balls) and drag them to its nest. Squee was described as  “75% reliable,” but it worked well only in a very dark room. Squee  was conceived by computer pioneer Edmund Berkeley, who earlier  wrote the hugely popular book Giant Brains or Machines That 


Think (1949). The original Squee prototype is in the permanent  collection of the Computer History Museum. (1,2) 

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. (3) 

1952 Arthur Samuel (IBM) wrote the first game-playing program,[35] for checkers (draughts), to achieve  sufficient skill to challenge a respectable amateur. His first checkers-playing program was written in  1952, and in 1955 he created a version that learned to play. (3) 

August 31, 1955 The term “artificial intelligence” is coined in a proposal for  a “2 month, 10 man study of artificial intelligence” submitted by John McCarthy  (Dartmouth College), Marvin Minsky (Harvard University), Nathaniel Rochester  (IBM), and Claude Shannon (Bell Telephone Laboratories). The workshop, which 


took place a year later, in July and August 1956, is generally considered as the  official birthdate of the new field. (1,2) 

1952: Arthur Samuel, a computer scientist, developed a checkers-playing  computer program – the first to independently learn how to play a game. (3) 

1955: John McCarthy and a team of men created a proposal for a workshop on  “artificial intelligence.” In 1956 when the workshop took place, the official birth of  the word was attributed to McCarthy. (3) 

December 1955 Herbert Simon and Allen Newell develop the Logic  Theorist, the first artificial intelligence program, which eventually would prove 38  of the first 52 theorems in Whitehead and Russell’s Principia Mathematica. Allen Newell, Herbert A. Simon and J.C. Shaw begin work on Logic  Theorist, a program that would eventually prove 38 theorems from  Whitehead and Russell’s Principia Mathematica. Logic  Theorist introduced several critical concepts to artificial intelligence  including heuristics, list processing and ‘reasoning as search.’ (1)


1955 The Turing test: Alan Turing creates a standard test to  answer: “Can machines think?” He proposed that if a computer, on  the basis of written replies to questions, could not be distinguished  from a human respondent, then it must be “thinking”. (1) 

1956 Robby the Robot: Robby the Robot appears in MGM’s 1956  science fiction movie Forbidden Planet. In the film, Robby was the  creation of Dr. Mobius and was built to specifications found in an  

alien computer system. Robby’s duties included assisting the  human crew while following Isaac Asimov’s Three Laws of Robotics  (1941). The movie was a cult hit, in part because of Robby’s 


humorous personality and Robby the Robot toys became huge  sellers. (3) 


The Dartmouth College summer AI conference is organized by John McCarthy, Marvin  Minsky, Nathan Rochester of IBM and Claude Shannon. McCarthy coins the term artificial  intelligence for the conference. (1) 

  • The first demonstration of the Logic Theorist (LT) written by Allen Newell, J.C. Shaw and Herbert A. Simon (Carnegie Institute of Technology, now Carnegie Mellon University or CMU). This is often called the first AI program, though Samuel’s checkers program also has a strong claim. See Over the holidays 50 years ago, two scientists hatched artificial intelligence. (2)


1957 Frank Rosenblatt develops the Perceptron, an early  artificial neural network enabling pattern recognition based on a two-layer  computer learning network. The New York Times reported the Perceptron to be  “the embryo of an electronic computer that [the Navy] expects will be able to  walk, talk, see, write, reproduce itself and be conscious of its existence.” The New  Yorker called it a “remarkable machine… capable of what amounts to thought.” (2) 

1957 The General Problem Solver (GPS) demonstrated by Newell, Shaw & Simon. (1) 

1958 John McCarthy (Massachusetts Institute of Technology or MIT) develops programming language Lisp which becomes the most popular  programming language used in artificial intelligence research. The  programming language LISP (short for “List Processing”) is invented in 1958 by John McCarthy at MIT. A key feature of LISP was that  data and programs were simply lists in parentheses, allowing a  program to treat another program – or itself – as data. This  characteristic greatly eased the kind of programming that attempted to model human thought. LISP is still used in a large number of  artificial intelligence applications.(1)



Herbert Gelernter and Nathan Rochester (IBM) described a theorem prover in geometry that exploits a  semantic model of the domain in the form of diagrams of “typical” cases. (1) 

Teddington Conference on the Mechanization of Thought Processes was held in the UK and among the  papers presented were John McCarthy’s Programs with Common Sense, Oliver  Selfridge‘s Pandemonium, and Marvin Minsky‘s Some Methods of Heuristic Programming and Artificial  Intelligence. (1) 

1959 The General Problem Solver (GPS) was created by Newell, Shaw and Simon while at CMU. (1) 1959 John McCarthy and Marvin Minsky founded the MIT AI Lab. (2) 

1959 Arthur Samuel coins the term “machine  learning,” reporting on programming a computer “so that it will learn to play a  better game of checkers than can be played by the person who wrote the  program.” (1) 

1959 Oliver Selfridge publishes “Pandemonium: A paradigm for  learning” in the Proceedings of the Symposium on Mechanization of Thought 


Processes, in which he describes a model for a process by which computers could  recognize patterns that have not been specified in advance. (1,2) 

1959 John McCarthy publishes “Programs with Common  Sense” in the Proceedings of the Symposium on Mechanization of Thought  Processes, in which he describes the Advice Taker, a program for solving  problems by manipulating sentences in formal languages with the ultimate  objective of making programs “that learn from their experience as effectively as  humans do.” (1) 

1959 Automatically Programmed Tools (APT): MIT´s  

Servomechanisms Laboratory demonstrates computer assisted  manufacturing (CAM). The school´s Automatically Programmed  Tools project created a language, APT, used to control milling  machine operations. At the demonstration, an air force general  claimed that the new technology would enable the United States to  “build a war machine that nobody would want to tackle.” The  machine produced a commemorative ashtray for each attendee. (3)




  • Arthur Samuel (IBM) wrote the first game-playing program, for checkers, to achieve sufficient skill to challenge a world champion. Samuel’s machine learning programs were responsible for the high performance of the checkers player. (3) 

AI in the 1960s 

Innovation in the field of artificial intelligence grew rapidly through the 1960s. The  creation of new programming languages, robots and automatons, research  studies, and films that depicted artificially intelligent beings increased in  popularity. This heavily highlighted the importance of AI in the second half of the  20th century.


Late 50’s & Early 60’s 

  • Margaret Masterman & colleagues at Cambridge design semantic nets for machine translation. See Themes in the work of Margaret Masterman by Yorick Wilks (1988). (2)
  • Ray Solomonoff lays the foundations of a mathematical theory of AI, introducing universal Bayesian methods for inductive inference and prediction. (1)
  • Man-Computer Symbiosis by J.C.R. Licklider. (2)

1960 Quicksort Algorithm: While studying machine translation of  languages in Moscow, C. A. R. Hoare develops Quicksort, an  algorithm that would become one of the most used sorting methods  in the world. Later, Hoare went to work for the British computer company Elliott Brothers, where he designed the first commercial  Algol 60 compiler. Queen Elizabeth II knighted C.A.R. Hoare in  2000. (1) 

1961 James Slagle (PhD dissertation, MIT) wrote (in Lisp) the first symbolic integration program,  SAINT, which solved calculus problems at the college freshman level. (1) 

1961 In Minds, Machines and Gödel, John Lucas[38] denied the possibility of machine intelligence  on logical or philosophical grounds. He referred to Kurt Gödel‘s result of 1931: sufficiently powerful  formal systems are either inconsistent or allow for formulating true theorems unprovable by any theorem-


proving AI deriving all provable theorems from the axioms. Since humans are able to “see” the truth of  such theorems, machines were deemed inferior. (1) 

1961: Unimate, an industrial robot invented by George Devol in the 1950s,  became the first to work on a General Motors assembly line in New Jersey. Its  responsibilities included transporting die castings from the assembly line and  welding the parts on to cars – a task deemed dangerous for humans. UNIMATE, the first mass-produced industrial robot, begins work at  General Motors. Obeying step-by-step commands stored on a  magnetic drum, the 4,000-pound robot arm sequenced and stacked  hot pieces of die-cast metal. UNIMATE was the brainchild of Joe  Engelberger and George Devol, and originally automated the  manufacture of TV picture tubes. (3)


1961 James Slagle develops SAINT (Symbolic Automatic  INTegrator), a heuristic program that solved symbolic integration problems in  freshman calculus. (1) 

1961 In Minds, Machines and Gödel, John Lucas[38] denied the possibility of machine intelligence  on logical or philosophical grounds. He referred to Kurt Gödel‘s result of 1931: sufficiently powerful  formal systems are either inconsistent or allow for formulating true theorems unprovable by any theorem proving AI deriving all provable theorems from the axioms. Since humans are able to “see” the truth of  such theorems, machines were deemed inferior. (1) 


  • First industrial robot company, Unimation, founded.

1963 The Ranchor Arm: Researchers design the Rancho Arm robot  at Rancho Los Amigos Hospital in Downey, California as a tool for  the handicapped. The Rancho Arm´s six joints gave it the flexibility  of a human arm. Acquired by Stanford University in 1963, it holds a  place among the first artificial robotic arms to be controlled by a  computer. (3)



  • Thomas Evans’ program, ANALOGY, written as part of his PhD work at MIT, demonstrated that computers can solve the same analogy problems as are given on IQ tests. (1)
  • Edward Feigenbaum and Julian Feldman published Computers and Thought, the first collection of articles about artificial intelligence.(1)
  • Leonard Uhr and Charles Vossler published “A Pattern Recognition Program That Generates, Evaluates, and Adjusts Its Own Operators”, which described one of the first machine learning programs that could adaptively acquire and modify features and thereby overcome the limitations  of simple perceptrons of Rosenblatt. (1) 
  • Ivan Sutherland’s MIT dissertation on Sketchpad introduced the idea of interactive graphics into computing. (1)


  • Edward A. Feigenbaum & Julian Feldman published Computers and Thought, the first collection of articles about artificial intelligence. (1).


Bertram Raphael‘s MIT dissertation on the SIR program demonstrates the power of a logical  representation of knowledge for question-answering systems. (1) 

1964 Daniel Bobrow completes his MIT PhD dissertation titled  Natural Language Input for a Computer Problem Solving System” and develops  STUDENT, a natural language understanding computer program. (1) 

1965 Lotfi Zadeh at U.C. Berkeley publishes his first paper introducing fuzzy logic “Fuzzy Sets”  (Information and Control 8: 338–353). (1) 

1965 J. Alan Robinson invented a mechanical proof procedure, the Resolution Method, which allowed  programs to work efficiently with formal logic as a representation language. (See Carl Hewitt’s  downloadable PDF file Middle History of Logic Programming). (2) 

1965 Joseph Weizenbaum (MIT) built ELIZA, an interactive program that carries on a dialogue  in English language on any topic. It was a popular toy at AI centers on the ARPANET when a version  that “simulated” the dialogue of a psychotherapist was programmed. (3) 

1965 Edward Feigenbaum initiated Dendral, a ten-year effort to develop software to deduce the  molecular structure of organic compounds using scientific instrument data. It was the first expert system.  (3) 

1965 DENDRAL artificial program: A Stanford team led by  professors Ed Feigenbaum, Joshua Lederberg and Carl Djerassi  creates DENDRAL, the first “expert system.” DENDRAL was an  artificial intelligence program designed to apply the accumulated  expertise of specialists to problem solving. Its area of specialization  was chemistry and physics. It applied a battery of “if-then” rules to  identify the molecular structure of organic compounds, in some  cases more accurately than experts. (1,2)


1965 Herbert Simon predicts that “machines will be capable,  within twenty years, of doing any work a man can do.” (1) 

1965 Hubert Dreyfus publishes “Alchemy and AI,” arguing that  the mind is not like a computer and that there were limits beyond which AI would  not progress. (1) 

1965 I.J. Good writes in “Speculations Concerning the First  Ultraintelligent Machine” that “the first ultraintelligent machine is the last 


invention that man need ever make, provided that the machine is docile enough  to tell us how to keep it under control.” (1) 

1965 Edward Feigenbaum, Bruce G. Buchanan, Joshua  Lederberg, and Carl Djerassi start working on DENDRAL at Stanford University.  The first expert system, it automated the decision-making process and problem solving behavior of organic chemists, with the general aim of studying hypothesis  formation and constructing models of empirical induction in science. Dendral program (Edward Feigenbaum, Joshua Lederberg, Bruce Buchanan, Georgia Sutherland  at Stanford University) demonstrated to interpret mass spectra on organic chemical compounds. First  successful knowledge-based program for scientific reasoning. (1,2) 

1965: Joseph Weizenbaum, computer scientist and professor, developed ELIZA,  an interactive computer program that could functionally converse in English with a person. It was a popular toy at AI centers on the ARPA-net when a version  that “simulated” the dialogue of a psychotherapist was programmed. Weizenbaum’s goal was to demonstrate how communication between an  artificially intelligent mind versus a human mind was “superficial,” but discovered  many people attributed anthropomorphic characteristics to ELIZA. ELIZA, an  interactive program that carries on a dialogue in English language on any topic.  Weizenbaum, who wanted to demonstrate the superficiality of communication  between man and machine, was surprised by the number of people who  attributed human-like feelings to the computer program. Joseph 



  • Ross Quillian (PhD dissertation, Carnegie Inst. of Technology, now CMU)

demonstrated semantic nets. (1) 

  • Machine Intelligence[39] workshop at Edinburgh – the first of an influential annual series organized by Donald Michie and others. (2)
  • Negative report on machine translation kills much work in Natural language processing (NLP) for many years. (3)

1966 Joseph Weizenbaum’s ELIZA : Joseph Weizenbaum finishes  ELIZA. ELIZA is a natural language processing environment. Its  most famous mode was called DOCTOR, which responded to user  questions much like a psychotherapist. DOCTOR was able to trick 


some users into believing they were interacting with another human,  at least until the program reached its limitations and became  nonsensical. DOCTOR used predetermined phrases or questions  and would substitute key words to mimic a human actually listening  to user queries or statements. (3) 

1966 The ORM: Developed at Stanford University, the Orm robot  (Norwegian for “snake”) was an unusual air-powered robotic arm. It  moved by inflating one or more of its 28 rubber bladders that were  sandwiched between seven metal disks. The design was  abandoned because movements could not be repeated accurately.



1966 Shakey the robot is the first general-purpose mobile  robot to be able to reason about its own actions. In a Life magazine 1970  article about this “first electronic person,” Marvin Minsky is quoted saying with  “certitude”: “In from three to eight years we will have a machine with the general  intelligence of an average human being.”. Shakey the Robot, developed by  Charles Rosen with the help of 11 others, was the first general-purpose mobile  robot, also known as the “first electronic person.” (3) 

Late 60s 

  1. Doug Engelbart invented the mouse at SRI.



Joel Moses (PhD work at MIT) demonstrated the power of symbolic reasoning for integration problems in  the Macsyma program. First successful knowledge-based program in mathematics. (2) 

Richard Greenblatt (programmer) at MIT built a knowledge-based chess-playing program, MacHack, that  was good enough to achieve a class-C rating in tournament play. (3) 

Wallace and Boulton’s program, Snob (Comp.J. 11(2) 1968), for unsupervised classification (clustering)  uses the Bayesian Minimum Message Length criterion, a mathematical realisation of Occam’s razor. (3) 

Marvin Minsky & Seymour Papert publish Perceptrons, demonstrating limits of simple neural nets. (1) 

1968: The sci-fi film 2001: A Space Odyssey (2001: Space Odyssey), directed by  Stanley Kubrick, is released. It features HAL (Heuristically programmed  ALgorithmic computer), a sentient computer. HAL controls the spacecraft’s  systems and interacts with the ship’s crew, conversing with them as if HAL were  human until a malfunction changes HAL’s interactions in a negative manner. (2)


1968 SHRDLU natural language: Terry Winograd begins work on  his PhD thesis at MIT. His thesis focused on SHRDLU, a natural  language used in artificial intelligence research. While precursor  programs like ELIZA were incapable of truly understanding English  commands and responding appropriately, SHRDLU was able to  combine syntax, meaning and deductive reasoning to accomplish  this. SHRDLU’s universe was also very simple, and commands  consisted of picking up and moving blocks, cones and pyramids of  various shapes and colors. Terry Winograd develops SHRDLU, an early  natural language understanding computer program. (1)



Stanford Research Institute (SRI): Shakey the Robot, demonstrated combining animal  locomotion, perception and problem solving. (3) 

Roger Schank (Stanford) defined conceptual dependency model for natural language  understanding. Later developed (in PhD dissertations at Yale University) for use in story  understanding by Robert Wilensky and Wendy Lehnert, and for use in understanding memory by  Janet Kolodner. (1) 

Yorick Wilks (Stanford) developed the semantic coherence view of language called Preference  Semantics, embodied in the first semantics-driven machine translation program, and the basis of  many PhD dissertations since such as Bran Boguraev and David Carter at Cambridge. (1) 

First International Joint Conference on Artificial Intelligence (IJCAI) held at Stanford. (2)



Marvin Minsky and Seymour Papert publish Perceptrons, demonstrating previously unrecognized  limits of this feed-forward two-layered structure, and This book is considered by some to mark  the beginning of the AI winter of the 1970s, a failure of confidence and funding for AI.  Nevertheless, significant progress in the field continued (see below). (1) 

McCarthy and Hayes started the discussion about the frame problem with their essay, “Some  Philosophical Problems from the Standpoint of Artificial Intelligence”. (1) 

1969 Arthur Bryson and Yu-Chi Ho  

describe backpropagation as a multi-stage dynamic system optimization method.  A learning algorithm for multi-layer artificial neural networks, it has contributed  significantly to the success of deep learning in the 2000s and 2010s, once  computing power has sufficiently advanced to accommodate the training of large  networks. (1,2) 

1969 Marvin Minsky and Seymour Papert publish Perceptrons:  An Introduction to Computational Geometry, highlighting the limitations of  simple neural networks. In an expanded edition published in 1988, they  responded to claims that their 1969 conclusions significantly reduced funding for  neural network research: “Our version is that progress had already come to a  virtual halt because of the lack of adequate basic theories… by the mid-1960s  there had been a great many experiments with perceptrons, but no one had been  able to explain why they were able to recognize certain kinds of patterns and not  others.” (1) 

1969 Victor Scheinmen’s Stanford arm: Victor Scheinman´s Stanford  Arm robot makes a breakthrough as the first successful electrically  powered, computer-controlled robot arm. By 1974, the Stanford  Arm could assemble a Ford Model T water pump, guiding itself with  optical and contact sensors. The Stanford Arm led directly to  commercial production. Scheinman then designed the PUMA series 


of industrial robots for Unimation, robots used for automobile  assembly and other industrial tasks. (3) 


1969 The Tentacle arm: Marvin Minsky develops the  Tentacle Arm robot, which moves like an octopus. It has twelve  joints designed to reach around obstacles. A DEC PDP-6 computer  controls the arm, powered by hydraulic fluids. Mounted on a wall, it  could lift the weight of a person. (3)


AI in the 1970s 

Like the 1960s, the 1970s gave way to accelerated advancements, particularly  focusing on robots and automatons. However, artificial intelligence in the 1970s  faced challenges, such as reduced government support for AI research. 

Early 70’s 

Jane Robinson and Don Walker established an influential Natural Language Processing group at SRI. (3)


Seppo Linnainmaa publishes the reverse mode of automatic differentiation. This method became later  known as backpropagation, and is heavily used to train artificial neural networks. (1) 

Jaime Carbonell (Sr.) developed SCHOLAR, an interactive program for computer assisted  instruction based on semantic nets as the representation of knowledge. (3) 

Bill Woods described Augmented Transition Networks (ATN’s) as a representation for natural language  understanding. (1) 

Patrick Winston‘s PhD program, ARCH, at MIT learned concepts from examples in the world of children’s  blocks. (3) 

1970 Shakey the robot: SRI International´s Shakey robot becomes  the first mobile robot controlled by artificial intelligence. Equipped  with sensing devices and driven by a problem-solving program  called STRIPS, the robot found its way around the halls of SRI by  applying information about its environment to a route. Shakey used  a TV camera, laser range finder, and bump sensors to collect data,  which it then transmitted to a DEC PDP-10 and PDP-15. The  computer sent commands to Shakey over a radio link. Shakey could  move at a speed of 2 meters per hour. (3)


1970 The first anthropomorphic robot, the WABOT-1, is built at  Waseda University in Japan. It consisted of a limb-control system, a vision system  and a conversation system. (3)


1971 Terry Winograd‘s PhD thesis (MIT) demonstrated the ability of computers to understand English  sentences in a restricted world of children’s blocks, in a coupling of his language understanding  program, SHRDLU, with a robot arm that carried out instructions typed in English. (3) 

1971 Work on the Boyer-Moore theorem prover started in Edinburgh.[40] (1) 

1972 Prolog programming language developed by Alain Colmerauer. (1) 

1972 Earl Sacerdoti developed one of the first hierarchical planning programs, ABSTRIPS.(1) 

1972 LUNAR natural language information retrieval system: LUNAR, a  natural language information retrieval system is completed by  William Woods, Ronal Kaplan and Bonnie Nash-Webber at Bolt, 


Beranek and Newman (BBN). LUNAR helped geologists access,  compare and evaluate chemical-analysis data on moon rock and  soil composition from the Apollo 11 mission. Woods was the  manager of the BBN AI Department throughout the 1970s and into  the early 1980s. (3)


1972 MYCIN, an early expert system for identifying bacteria  causing severe infections and recommending antibiotics, is developed at Stanford  University. (1) 

1973 James Lighthill reports to the British Science Research  Council on the state artificial intelligence research, concluding that “in no part of  the field have discoveries made so far produced the major impact that was then  promised,” leading to drastically reduced government support for AI research. (3) 

1973 The Assembly Robotics Group at University of Edinburgh builds Freddy Robot, capable of  using visual perception to locate and assemble models. (See Edinburgh Freddy Assembly Robot: a  versatile computer-controlled assembly system.) (3) 

1973: James Lighthill, applied mathematician, reported the state of artificial  intelligence research to the British Science Council, stating: “in no part of the field  have discoveries made so far produced the major impact that was then  promised,” The Lighthill report gives a largely negative verdict on AI research in Great Britain and  forms the basis for the decision by the British government to discontinue support for AI research in all but  two universities. (1) 

1973: Inventing the Internet, TCP/IP internetworking protocol, first sketched  out in 1973 by Vint Cerf and Bob Kahn. (1, 3). 

1974 The silver arm: David Silver at MIT designs the Silver Arm, a  robotic arm to do small-parts assembly using feedback from  delicate touch and pressure sensors. The arm´s fine movements  approximate those of human fingers. (3)



Ted Shortliffe‘s PhD dissertation on the MYCIN program (Stanford) demonstrated a very practical rule based approach to medical diagnoses, even in the presence of uncertainty. While it borrowed from  DENDRAL, its own contributions strongly influenced the future of expert system development, especially  commercial systems. (1) 

Earl Sacerdoti developed one of the first planning programs, ABSTRIPS, and developed techniques of  hierarchical planning. (1) 


Earl Sacerdoti developed techniques of partial-order planning in his NOAH system, replacing the  previous paradigm of search among state space descriptions. NOAH was applied at SRI International to  interactively diagnose and repair electromechanical systems. (3) 

Austin Tate developed the Nonlin hierarchical planning system able to search a space of partial  plans characterised as alternative approaches to the underlying goal structure of the plan.(2) 

Marvin Minsky published his widely read and influential article on Frames as a representation of  knowledge, in which many ideas about schemas and semantic links are brought together. (1) 

Marvin Minsky published his widely read and influential article on Frames as a representation of  knowledge, in which many ideas about schemas and semantic links are brought together. (1) 

The Meta-Dendral learning program produced new results in chemistry (some rules of mass  spectrometry) the first scientific discoveries by a computer to be published in a refereed journal. (1) 

Mid 70’s 

  1. Barbara Grosz (SRI) established limits to traditional AI approaches to discourse modeling. Subsequent work by Grosz, Bonnie Webber and Candace Sidner developed the notion of “centering”, used in establishing focus of discourse and anaphoric references in Natural  language processing. (1,3) 
  2. David Marr and MIT colleagues describe the “primal sketch” and its role in visual perception. (1)
  3. Alan Kay and Adele Goldberg (Xerox PARC) developed the Smalltalk language, establishing the power of object-oriented programming and of icon-oriented interfaces.
  4. David Marr and MIT colleagues describe the “primal sketch” and its role in visual perception.


  1. Douglas Lenat‘s AM program (Stanford PhD dissertation) demonstrated the discovery model (loosely guided search for interesting conjectures). (1)


  1. Randall Davis demonstrated the power of meta-level reasoning in his PhD dissertation at Stanford. (1)

1976 Shigeo Hirose’s Soft Gripper: Shigeo Hirose´s Soft Gripper  robot can conform to the shape of a grasped object, such as a wine  glass filled with flowers. The design Hirose created at the Tokyo  Institute of Technology grew from his studies of flexible structures in  nature, such as elephant trunks and snake spinal cords. (3)


1976 Computer scientist Raj Reddy publishes “Speech  Recognition by Machine: A Review” in the Proceedings of the IEEE, summarizing  the early work on Natural Language Processing (NLP). (1) 

1977: Director George Lucas’ film Star Wars is released. The film features C 3PO, a humanoid robot who is designed as a protocol droid and is “fluent in more  than seven million forms of communication.” As a companion to C-3PO, the film  also features R2-D2 – a small, astromech droid who is incapable of human  speech (the inverse of C-3PO); instead, R2-D2 communicates with electronic  beeps (2)




Its functions include small repairs and co-piloting starfighters. 

Late 70’s 

Stanford’s SUMEX-AIM resource, headed by Ed Feigenbaum and Joshua Lederberg,  demonstrates the power of the ARPAnet for scientific collaboration. (3) 


Tom Mitchell, at Stanford, invented the concept of Version spaces for describing the search  space of a concept formation program. (2) 

Herbert A. Simon wins the Nobel Prize in Economics for his theory of bounded rationality, one of  the cornerstones of AI known as “satisficing“. (1) 

The MOLGEN program, written at Stanford by Mark Stefik and Peter Friedland, demonstrated  that an object-oriented programming representation of knowledge can be used to plan gene cloning experiments. (1) 

1978 The XCON (eXpert CONfigurer) program, a rule-based  expert system assisting in the ordering of DEC’s VAX computers by automatically  selecting the components based on the customer’s requirements, is developed at  Carnegie Mellon University. 

1978 Speak & Spell: Texas Instruments Inc. introduces  Speak & Spell, a talking learning aid for children aged 7 and up. Its  debut marked the first electronic duplication of the human vocal  tract on a single integrated circuit. Speak & Spell used linear  predictive coding to formulate a mathematical model of the human  vocal tract and predict a speech sample based on previous input. It  transformed digital information processed through a filter into  synthetic speech and could store more than 100 seconds of  linguistic sounds. (3)





Mycin program, initially written as Ted Shortliffe’s Ph.D. dissertation at Stanford, was  demonstrated to perform at the level of experts. Bill VanMelle’s PhD dissertation at Stanford  demonstrated the generality of MYCIN’s representation of knowledge and style of reasoning in  his EMYCIN program, the model for many commercial expert system “shells”. (1) 

Jack Myers and Harry Pople at University of Pittsburgh developed INTERNIST, a knowledge based medical diagnosis program based on Dr. Myers’ clinical knowledge. (3) 

Cordell Green, David Barstow, Elaine Kant and others at Stanford demonstrated the CHI system  for automatic programming. (3)



BKG, a backgammon program written by Hans Berliner at CMU, defeats the reigning world  champion (in part via luck). (1) 

Drew McDermott and Jon Doyle at MIT, and John McCarthy at Stanford begin publishing work  on non-monotonic logics and formal aspects of truth maintenance.(1) 

1979: The Stanford Cart, a remote controlled, tv-equipped mobile robot was  created by then- mechanical engineering grad student James L. Adams in 1961.  In 1979, a “slider,” or mechanical swivel that moved the TV camera from side-to side, was added by Hans Moravec, then-PhD student. The cart successfully  crossed a chair-filled room without human interference in approximately five  hours, making it one of the earliest examples of an autonomous vehicle. The  Stanford Cart was a long-term research project undertaken at  Stanford University between 1960 and 1980. In 1979, it successfully  crossed a room on its own while navigating around a chair placed  as an obstacle. Hans Moravec rebuilt the Stanford Cart in 1977,  equipping it with stereo vision. A television camera, mounted on a  rail on the top of the cart, took pictures from several different angles  and relayed them to a computer. The Stanford Cart, built by Hans Moravec, becomes  the first computer-controlled, autonomous vehicle when it successfully traverses a chair-filled room and  circumnavigates the Stanford AI Lab. The Stanford Cart successfully crosses a chair-filled  room without human intervention in about five hours, becoming one of the  earliest examples of an autonomous vehicle. (3)




AI in the 1980s 

The rapid growth of artificial intelligence continued through the 1980s. Despite  advancements and excitement behind AI, caution surrounded an inevitable “AI  Winter,” a period of reduced funding and interest in artificial intelligence. 

Lisp machines developed and marketed. First expert system shells and commercial applications. (1,3) First expert system shells and commercial applications. (3) 

1980 First National Conference of the American Association for Artificial Intelligence (AAAI) held at  Stanford. (3) 

1980: WABOT-2 (Wabot-2) was built at Waseda University. This inception of the  WABOT allowed the humanoid to communicate with people as well as read  musical scores and play tunes of average difficulty on an electronic organ. (3)





Lee Erman, Rick Hayes-Roth, Victor Lesser and Raj Reddy published the first description of the  blackboard model, as the framework for the HEARSAY-II speech understanding system. (1) 

First National Conference of the American Association of Artificial Intelligence (AAAI) held at  Stanford. (3) 

1981 Danny Hillis designs the connection machine, which utilizes Parallel computing to bring new  power to AI, and to computation in general. (Later founds Thinking Machines Corporation) (3) 

1981 The Japanese Ministry of International Trade and  Industry budgets $850 million for the Fifth Generation Computer project. The  project aimed to develop computers that could carry on conversations, translate  languages, interpret pictures, and reason like human beings. (3) 

1981 The direct drive arm: The first direct drive (DD) arm by Takeo  Kanade serves as the prototype for DD arms used in industry today.  The electric motors housed inside the joints eliminated the need for  the chains or tendons used in earlier robots. DD arms were fast and  accurate because they minimize friction and backlash. (3)


  The Boston Global Forum * The Michael Dukakis Institute for Leadership and Innovation 

AI World Society Innovation Network,, [email protected], 67 Mount Vernon St, Unit F, Boston, MA 02108 


1982 The FRED: Nolan Bushnell founded Androbot with former  Atari engineers to make playful robots. The “Friendly Robotic  Educational Device” (FRED), designed for 6-15 year-olds, never  made it to market. (3) 

1982 The Fifth Generation Computer Systems project (FGCS), an initiative by Japan’s Ministry of  International Trade and Industry, begun in 1982, to create a “fifth generation computer” (see history of  computing hardware) which was supposed to perform much calculation utilizing massive parallelism. (3) 




John Laird and Paul Rosenbloom, working with Allen Newell, complete CMU dissertations  on Soar (program). (1) 

James F. Allen invents the Interval Calculus, the first widely used formalization of temporal events. (2) 

1984: The film Electric Dreams, directed by Steve Barron, is released. The plot  revolves around a love triangle between a man, woman, and a sentient personal  computer called “Edgar.” (2) 

1984: At the Association for the Advancement of Artificial Intelligence (AAAI),  Roger Schank (AI theorist) and Marvin Minsky (cognitive scientist) warn of the “AI  Winter”, predicting an immanent bursting of the AI bubble (which did happen 



three years later), similar to the reduction in AI investment and research funding  in the mid-1970s. The first instance where interest and funding for artificial  intelligence research would decrease. (3) 

1984 Electric Dreams is released, a film about a love triangle  between a man, a woman and a personal computer. (2) 

1984 The IBM 7535: Based on a Japanese robot, IBM’s 7535 was  controlled by an IBM PC and programmed in IBM’s AML (“A Manufacturing Language”). It could manipulate objects weighing up  to 13 pounds. (3)

1984 Hero Jr. robot kit: Heathkit introduces the Hero Jr. home  robot kit, one of several robots it sells at the time. Hero Jr. could  roam hallways guided by sonar, play games, sing songs and even  act as an alarm clock. The brochure claimed it “seeks to remain  near human companions” by listening for voices. (3)


Mid 80’s 

Neural Networks become widely used with the Backpropagation algorithm algorithm (first described by  Werbos in 1974), also known as the reverse mode of automatic differentiation published by Seppo  Linnainmaa in 1970 and applied to neural networks by Paul Werbos

1985 The autonomous drawing program, AARON, created by Harold Cohen, is demonstrated at the  AAAI National Conference (based on more than a decade of work, and with subsequent work showing  major developments). 

1985 Denning Sentry Robot: Boston-based Denning designed the  Sentry robot as a security guard patrolling for up to 14 hours at 3  mph. It radioed an alert about anything unusual in a 150-foot radius.  The product, and the company, did not succeed.

1986 First driverless car, a Mercedes-Benz van equipped with  cameras and sensors, built at Bundeswehr University in Munich under the  direction of Ernst Dickmanns, drives up to 55 mph on empty streets. 

1986 Barbara Grosz and Candace Sidner create the first computation model of discourse, establishing  the field of research.

1986 LMI Lambda: The LMI Lambda LISP workstation is  introduced. LISP, the preferred language for AI, ran slowly on  expensive conventional computers. This specialized LISP computer,  both faster and cheaper, was based on the CADR machine  designed at MIT by Richard Greenblatt and Thomas Knight. 

1986 Omnibot 2000: The Omnibot 2000 remote-controlled  programmable robot toy could move, talk and carry objects. The  cassette player in its chest recorded actions to be taken and  speech to be played. 

October 1986 David Rumelhart, Geoffrey Hinton, and Ronald Williams  publish ”Learning representations by back-propagating errors,” in which they  describe “a new learning procedure, back-propagation, for networks of neurone like units.” 

1987 Mitsubisi Movemaster RM-501 Gripper is introduced: The  Mitsubishi Movemaster RM-501 Gripper is introduced. This robot  gripper and arm was a small, commercially available industrial  robot. It was used for tasks such as assembling products or  handling chemicals. The arm, including the gripper, had six degrees  of freedom and was driven by electric motors connected to the  joints by belts. The arm could move fifteen inches per second, could  lift 2.7 pounds, and was accurate within .02 of an inch. 

1987 Marvin Minsky published The Society of Mind, a theoretical description of the mind as a collection  of cooperating agents. He had been lecturing on the idea for years before the book came out (c.f. Doyle  1983).


1987 Around the same time, Rodney Brooks introduced the subsumption architecture and behavior based robotics as a more minimalist modular model of natural intelligence; Nouvelle AI

1987 Commercial launch of generation 2.0 of Alacrity by Alacritous Inc./Allstar Advice Inc. Toronto, the  first commercial strategic and managerial advisory system. The system was based upon a forward chaining, self-developed expert system with 3,000 rules about the evolution of markets and competitive  strategies and co-authored by Alistair Davidson and Mary Chung, founders of the firm with the underlying  engine developed by Paul Tarvydas. The Alacrity system also included a small financial expert system  that interpreted financial statements and models. 

1987 The video Knowledge Navigator, accompanying Apple  CEO John Sculley’s keynote speech at Educom, envisions a future in which  “knowledge applications would be accessed by smart agents working over  networks connected to massive amounts of digitized information.” 

1988 Judea Pearl publishes Probabilistic Reasoning in  Intelligent Systems. His 2011 Turing Award citation reads: “Judea Pearl created  the representational and computational foundation for the processing of  information under uncertainty. He is credited with the invention of Bayesian  networks, a mathematical formalism for defining complex probability models, as  well as the principal algorithms used for inference in these models. This work not  only revolutionized the field of artificial intelligence but also became an important  tool for many other branches of engineering and the natural sciences.” 

1988 Rollo Carpenter develops the chat-bot Jabberwacky to  “simulate natural human chat in an interesting, entertaining and humorous  manner.” It is an early attempt at creating artificial intelligence through human  interaction. 

1988 Members of the IBM T.J. Watson Research Center publish  “A statistical approach to language translation,” heralding the shift from rule based to probabilistic methods of machine translation, and reflecting a broader shift to “machine learning” based on statistical analysis of known examples, not  comprehension and “understanding” of the task at hand (IBM’s project Candide,  successfully translating between English and French, was based on 2.2 million  pairs of sentences, mostly from the bilingual proceedings of the Canadian  parliament). 

1988 Marvin Minsky and Seymour Papert publish an expanded  edition of their 1969 book Perceptrons. In “Prologue: A View from 1988” they  wrote: “One reason why progress has been so slow in this field is that researchers  unfamiliar with its history have continued to make many of the same mistakes  that others have made before them.”

1988: Computer scientist and philosopher Judea Pearl published “Probabilistic  Reasoning in Intelligent Systems.” Pearl is also credited with inventing Bayesian  networks, a “probabilistic graphical model” that represents sets of variables and  their dependencies via directed acyclic graph (DAG). 

1988: Rollo Carpenter, programmer and inventor of two chatbots, Jabberwacky  and Cleverbot (released in the 1990s), developed Jabberwacky to “simulate  natural human chat in an interesting, entertaining and humorous manner.” This is  an example of AI via a chatbot communicating with people. 

1989 Yann LeCun and other researchers at AT&T Bell  Labs successfully apply a backpropagation algorithm to a multi-layer neural  network, recognizing handwritten ZIP codes. Given the hardware limitations at  the time, it took about 3 days (still a significant improvement over earlier efforts)  to train the network. 

Dean Pomerleau at CMU creates ALVINN (An Autonomous Land Vehicle in a Neural Network), which grew into the system that drove a car coast-to-coast under computer control for all butabout 50 of the 2850 miles. The development of metal–oxide–semiconductor (MOS) very-large-scale integration (VLSI), in the form of complementary MOS (CMOS) technology, enabled the development ofpractical artificial neural network (ANN) technology in the 1980s. A landmark publication in the field was the 1989 book Analog VLSI Implementation of Neural Systems by Carver A. Mead and Mohammed Ismail.

1989 Computer defeats master chess player: David Levy is the firstmaster chess player to be defeated by a computer. The program Deep Thought defeats Levy who had beaten all other previous computer counterparts since 1968.

AI in the 1990s
The end of the millennium was on the horizon, but this anticipation only helped artificial intelligence in its continued stages of growth.


Major advances in all areas of AI, with significant demonstrations in machine learning, intelligent tutoring, case-based reasoning, multi-agent planning, scheduling, uncertain reasoning, data mining, natural language understanding and translation, vision, virtual reality, games, and other topics.
Early 1990s TD-Gammon, a backgammon program written by Gerry Tesauro, demonstrates that reinforcement (learning) is powerful enough to create a championship-level game-playing program by competing favorably with world-class players.

Rod Brooks’ COG Project at MIT, with numerous collaborators, makes significant progress in building a humanoid robot EQP theorem prover at Argonne National Labs proves the Robbins Conjecture in mathematics (October-November, 1996). The Deep Blue chess program beats the current world chess champion, Garry Kasparov, in a widely followed match and rematch (See Deep Blue Wins). (May 11th, 1997). NASA’s pathfinder mission made a successful landing and the first autonomous robotics
system, Sojourner, was deployed on the surface of Mars. (July 4, 1997)
First official Robo-Cup soccer match (1997) featuring table-top matches with 40 teams of
interacting robots and over 5000 spectators.
Web crawlers and other AI-based information extraction programs become essential in widespread use of the world-wide-web.
Demonstration of an Intelligent Room and Emotional Agents at MIT’s AI Lab. Initiation of work on the Oxygen Architecture, which connects mobile and stationary computers in an adaptive network.