I Don’t Know Jack About Artificial Intelligence
Where does the story of Artificial Intelligence begin?
It begins with the building of small independent machines called automatons in ancient Greece. Automatons were deterministic machines that could do one or more tasks without any human intervention. For example, an automaton could start a fire, play a musical instrument by itself, move or “dance” independently, and even write meaningfully. The ancients used clockwork, steam, or water to make these automatons perform their functions.
Automatons were primarily used as curiosity objects to entertain royalty in ancient times. Nobody ever thought that one day they would gain abilities far greater than what they ever had.
The first recorded automaton to be ever built was the Wooden Pigeon of Archytas, created by the Greek philosopher Archytas in 400 BC.
The Greeks have several lores of machines and non-living human replicas coming to life. They speak of the God of the forge, Hephaestus, who built a giant living bronze automaton called Talos. Then there is the story of the inventor Dedalus, who built wax wings to fly with in order to escape from prison. He was the creator of other lifelike automata, which were believed by many to have some form of life. And then there is the sculptor Pygmalion who fell in love with a statue of his own making to whom the Goddess Aphrodite later granted life. He named his darling creature Galatea.
These were the nascent ideas of what is today known as Artificial Intelligence — attributing human-level intelligence to machines or non-living things. These ideas show that there has existed a strong fascination among people such as thinkers, writers, and scientists to give “birth” to or give “life” to something that can think, feel, and behave as humans do. That fascination is only now bearing fruit, however small.
Gulliver’s Travels, Ada Lovelace, Charles Babbage
In the 1700s, popular English satirist Jonathan Swift, imagined a certain machine in his famous book “Gulliver’s Travels”, named The Engine.
In brief, the Engine is a machine that randomly selects certain words on the throw of dice. The dice-throwing process is repeated until the random words string into a meaningful sentence, then it is read out and put into a book. This is how Swift imagined that books were written in the imaginary capital city of Lagado, a place visited by Gulliver, the eponymous hero of the story.
Strangely, this idea is probably the closest and most accurate depiction of today's Large Language Models like ChatGPT you will ever find on the other side of the 1900s.
No one had built such an engine before. The idea that such a machine, even one that could perform simple arithmetical operators, a tad similar to the abacus, was yet to be invented.
A 100 years later, through a man named Charles Babbage, the idea for a computer would occur to mankind.
The idea for the world’s first computer was inspired by the “punch cards” of the Jacquard loom.
The Jacquard loom was built by Joseph Marie Jacquard in 1801 to manufacture complex designs in cloth like damask and brocade.
The loom consisted of a chain of pasteboard cards with small holes in them. The presence or absence of a hole determined which of the warp threads would be raised and those which wouldn't, creating beautiful patterns and designs in the cloth.
Charles Babbage, an English mathematician and mechanical engineer, saw a picture of the Jacquard punch cards and it inspired him to use them as the source of input for a machine he proposed to build, the Analytical Engine.
Babbage had earlier made plans to build a Difference Engine, an early calculating machine, the precursor to the Analytical Engine — which was arguably a more ambitious project.
The Analytical Engine was never completed by Babbage. It was meant to perform any type of arithmetical calculation, but sadly, it never got made though he obtained grants from the British government for it.
Babbage’s mathematician friend Lady Ada Lovelace, the daughter of the famous Romantic poet Lord George Byron, was the first person to write a “program” that is a series of instructions instructing the Analytical Engine on how to make its calculations. Babbage and Lovelace’s work set the stage for the building of calculators or computers (machines which compute and calculate).
So far computers were mechanical and not electrical.
It’s only in 1946 that the first electrical computer ENIAC was built.
Robots, Gakutensoku, Turing
It’s noteworthy that during this period several writers gave expression to the creation of other living things or forms by humans.
During the Romantic Era, Mary Shelley wrote Frankenstein, the first-of-its-kind story about an inventor who created a botched up man-like monster, releasing him to his fate in the world. The monster was created by electricity.
In the 1900s, there was a lot of scientific debate about the creation of machines that could think.
Against that backdrop, in 1921, the Czech playwright Karel Čapek wrote a play where he introduced the concept of robots, a type of artificial people who would gain consciousness and kill their human overlords.
This play became very popular worldwide and inspired a scientist at the other end of the world, a Japanese professor of marine biology in the Japanese city of Sapporo to build a robot. He created the world’s first robot called Gakutensoku, which means “learning from the rules of nature.” This was in the year 1929.
Then came World War 2. The war brought with it the need to send secret messages from behind enemy lines by spies and intercept and quickly decipher the enemy’s messages.
Consequently encryption techniques and cryptography took off, which now brings us to a certain atheist and homosexual man by the name of Alan Turing, a genius in his own right who wrote a paper called “Computing Machinery and Intelligence” where he proposed the question “Can machines think?” and devised the Imitation Game as a means to show that if a machine sounds like it thinks, we can infer that it does. This set the stage for Artificial Intelligence as a legitimate field of study.
Simple Simon, The Imitation Game, Dartmouth Conference, Perceptron
We need to back up a bit now. During the Second World War, the Germans were the first to invent an electromechanical digital machine called the Z3 — the world’s first fully programmable computer. In response, the British invented Colossus, the world’s first programmable electronic computer that was used to break codes of the Lorenz SZ 40/42 machine, owned by the Germans.
During and after the war, several computers proliferated. These were the ENIAC, the Colossus Mark I and the Colossus Mark II.
A computer scientist, Edward Callis Berkeley, wrote a book “Giant Brains and Machines that Think” in 1949 where he talked about a simple computer named Simon that could perform only 4 calculations and knew only 4 numbers. However, what he describes in the book, about machines being able to think, caught my attention. Berkeley concedes that although we don't know much about how we think, there are some tasks which can be broken down into logical problems with a step-by-step process. It's in these tasks that computers really outsmart man. But in intuitive thinking, having brainwaves and impulsive thoughts, man outdoes the computer time and again. That is until recently. Now the primary work being done in Artificial Intelligence is to get the machine to recognise images and text, two inputs which were hitherto difficult for computers to work with.
In 1950, Alan Turing, the British hero responsible for cracking the German Enigma codes wrote a paper entitled "Computing Machinery and Intelligence". In this paper he proposed the Imitation Game now known as the Turing test.
Alan Turing did not actually answer the question “Can machines think?”, he simply mapped it to another problem which is known as The Imitation Game.
The Imitation Game goes thus:
There are 3 people involved. A man. A woman. An interrogator. All three are in separate rooms and communicate via intermediaries.
The role of the man is to deceive the interrogator into believing he is the woman.
The role of the woman is to help the interrogator know without a shadow of doubt that she is the woman.
The interrogator through questions alone must guess who is the man and who is the woman.
As you can see, in the game, one of them (the man) is trying to ape or imitate the other (the woman) to an extent where the neutral observer, the interrogator, cannot distinguish between the man and the woman at all from the answers they give to his questions.
Now, the issue is, the man in this case must outthink or outsmart the woman. He cannot be a mere spectator. He must be an ace fabricator to fool the interrogator into believing he is the woman. He must play the part of a devious creature by answering cunningly with the intent to deceive.
Now the woman herself must answer as truthfully as possible and with a touch of pleading to establish her identity.
The intelligence and capability of the interpreter is also something that must be considered here.
Turing next sought to replace the man with a machine and continue the game. Will the interrogator now be able to distinguish between the woman and the machine?
He mapped this problem to the now confounding question: “Can machines think?”
Ultimately the purpose of this experiment was to investigate whether a person could distinguish between a human and a machine, given that the machine spoke as if it could think. One isn’t interested in knowing whether the machine is capable of original thought, but whether it can appear to convey original or human-like thought. Because, and we come back to it, we still don’t know where original thought comes from. We’re clueless as to what intelligence is, what is thought, how we think and act.
Just by speaking like a human being, the computer can ape a person realistically. So it’s basically a copycat. But you may argue, isn’t that how children learn? By copying their mother’s speech? Are children capable of “original thinking” unless trained to do so? Similarly, the computers have now been trained to a degree, wherein their accuracy is child-like, they make mistakes just as children do (for which they are not as yet held accountable), they learn and better themselves. But it all started by aping. By copying. The next level is outsmarting.
In 1955, a proposal for the now-famous Dartmouth workshop entitled "A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence”, was submitted by John McCarthy of Dartmouth College, Marvin Minsky of Harvard University, Nathaniel Rochester from IBM and Claude Shannon from Bell Telephone Laboratories. The workshop is considered the birthplace of artificial intelligence. The aim of the workshop can be summed up in this one sentence: "every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it."
I’d like to pause here to tell the reader that the reason for summarising the history of the thought of Artificial Intelligence along with the growth of computers is this: one needs both the machinery and the algorithms to keep in step if progress is to be made in Artificial Intelligence. For example, it was only in the mid 2000s that the idea to use GPUs for AI computing came to us. So it’s necessary to understand how both human thought and the existing machinery can prove to be roadblocks in the growth of any new technology.
The outcomes of the Dartmouth Summer Research Project were manifold.
