The term “Artificial Intelligence” (AI) is not new. Since ancient times, dreamers, alchemists, artists, sculptors and scientists have been working tirelessly to convert the dream into concrete reality. Perhaps the first step, the Concept of AI, has been the most difficult, and taken the longest, to be frozen. But since this was first introduced in the 1940’s, the progress has been astonishing, and it seemed for a while that we were already at the doorstep. But recent reviews have proved otherwise. This article aims to give an accurate history of artificial intelligence from inception to current times.
The human brain is the most complex machine we know. In fact, our whole concept of Reality stems from the evaluation of our world by this one central organ. But we have been trying to understand how the Brain works from a very ancient time, recognizing that the seat of our power over the Earth, lay in here. The roots can be traced back to the classical philosophers of ancient Greece, among other centers of thinking on earth, such as the Chinese, Egyptians and Indians. It was the efforts by these Greek philosophers to model the mind as a series of symbols which connected to form a cogent system that started the ball rolling. The concept of “Humanoids” as human-like robots who could faithfully and intelligently serve Man without the inherent flaws of Humanity, has been the guide for much research into AI. Machines that can think more or less, like us, are the real goal for AI. Titled “Intelligibility”, this kind of intelligence can be used to explore our ability to learn and reason, to perceive and judge, and finally to execute mental actions and exercises.
The series of steps that culminated in the identification, titling and search for AI began in 1308, when Theologian Ramon and Llull the Catalan Poet published ‘Ars generalis ultima’ (The Ultimate General Art), which was a method using mechanical paper based talks about creating new familiarity from concepts from combinations. Following Llull, Mathematician Gottfried Leibniz in 1666 proposed a new Human Thought alphabet. He argued that every ideology is nothing but amalgamation of simpler concepts. In the mean time Science Fiction writers were also trying to conceptualize AI. In 1726, Jonathan Swift’s ‘Gulliver’s Travels’ describes a machine he calls The Engine, which can duplicate even the creative efforts of the human mind, and write books on philosophy, mathematics, theology, poetry and so on, without actual human assistance.
George Boole In 1854, created ‘Boolean Logic’ which argued that all logical thinking could be systematized as a set of mathematical equations. Then in Madison Square Garden at an Electrical exhibition in 1898, an extraordinary genius called Nicola Tesla demonstrated (a boat) the foremost radio-controlled vessel in the world. The boat was operational with what Tesla described as ‘A Borrowed Mind’.
Makoto Nishimura in 1929, designed the ‘Gakutensoku’ (‘learn from the laws of nature’, in Japanese), which was the initial robot fabricated in Japan. The facial expressions could be changed, and it also moved its head and hands.
Milestones – From where we are today to where it began
The “Logic Theorist” was one of the earliest AI programs for logic solving, and was written in 1956 to mimic the problem solving ability of the human mind. Tested against the historical “Principia Mathematica”, this program actually proved 38 of the first 52 theorems immediately. It even improved on some of these theorems in the process. For the first time, it was proved that machines could perform tasks that were considered till then to be in the specific domain of intelligent (and gifted) human beings only.
Inductive Thinking came next. Scientists use Inductive Thinking to create Hypothesis to explain the data being examined. In 1965, The Dendral Project was conducted by NASA Scientists to prove that the second feature of AI, “Instrumentality”, could be achieved by machines. The Dendral Project was based on creating a cognitive model which would help the NASA Scientists to identify organic molecules using machine knowledge of Organic Chemistry. AI was used to create a set of algorithms to accomplish molecule identification by inductive reasoning. It was the first time that a logic based set of ‘if/then’ rules were used alongside the cognitive model. This was the first example of an Expert System.
In 1950, the computer genius Allen Turing published ‘Computing Machinery and Intelligence’, in which his ‘imitation game’ (later called the ‘Turing Test’) was first proposed. Dean Edmunds and Marvin Minsky built (Stochastic Neural Analog Reinforcement Calculator) SNARC In 1951, which was built for network stimulation of Neurons with 3000 vacuum tubes, and was the world’s first Artificial Neural Network.
‘Artificial Intelligence’ was first proposed on August 31st 1955, as was a project for a two month, ten man study of AI. This proposal was submitted by Marvin Minsky, John McCarthy, Nathaniel Rochester & Claude Shannon, from various renowned universities and organizations. The actual workshop took place in 1956 July -August, and is the official birth-date of a new field of Artificial Intelligence.
In 1972, a premature Expert System to identify bacterium was first developed at Stanford University, causing severe infections and recommended suitable antibiotics.
Then in 2011, ‘Watson’, a natural language query answering super computer, based on AI, competes in Jeopardy! and defeats two former world champions.
In March 2016, Google DeepMind’s AlphaGo defeated Go Champion Lee Sedol, again using AI.
The Future of AI
The Future of AI is still not yet as bright as it should be. Despite the massive advances in Machine Learning and Robotics, the challenges for AI remain much the same as it was sixty years ago. The early experiments by cognitive scientists to determine “intelligibility” were done 60 years ago, and have not been bettered. The world of Hal, the sentient computer in Stanley Kubrick’s visionary film ‘2001: Space Odyssey’, is yet to be achieved.