Even though we live in an era in which we rely so much on iPhones, laptops, and the internet just to get through the day, there’s something about the phrase “Artificial Intelligence” that still seems curiously like science fiction instead of science fact. But the truth is, A.I. has been a very real thing in the world for more than a century. A.I.s have made for skilled chess players, therapists, and the most curiously compelling week of Jeopardy! ever broadcast. Here are four highlights from the history of our future robot overlords.
EL AJEDRECISTA: A Remarkable Automatic Device
In English, El Ajedrecista is “The Chess Player”—but, let’s be honest, “El Ajedrecista” sounds so much better. The machine astonished the world at the Paris World Fair of 1914 by using magnets and math to play an endgame of chess against a human opponent. The machine originally had mechanical arms to move the pieces, before the much less terrifying electromagnet solution was introduced.
El Ajedrecista may not have been the first time someone claimed to have invented a chess-playing automaton, but it was the first time that reality lived up to the hype; in 1769, Wolfgang von Kempelen’s “The Turk,” which made similar claims to El Ajedrecista, was later revealed as a hoax when a human player was discovered inside the machine. In contrast, El Ajedrecista was not only the first publicly displayed artificially-intelligent automaton, but it can also make a claim to being the world’s first computer game.
It’s fair to say that the results of each game were somewhat stacked in El Ajedrecista’s favor from the start. The machine does not play a full game of chess. It only works its way through an endgame, pitting its king and rook against the human opponent’s king, each starting from pre-set positions. The algorithm it uses cannot be beaten by any legal maneuver at that point in the game. Essentially, its creators just reverse-engineered the set-up, starting with checkmate and backing up to the last point at which the game becomes unwinnable unless the machine makes a mistake. (There’s no rule that robots have to have a sense of fair play.) El Ajedrecista is still on show—and still playable—in Madrid, and has never failed to defeat its opponent.
ELIZA: Mommy to a Million Chatbots
Considered the forerunner to the modern-day chatbot, ELIZA (created 1964) was the creation of Joseph Weizenbaum of the Massachusetts Institute of Technology, and ended up turning Weizenbaum from an early advocate of artificial intelligence into one of its harshest critics.
What was the problem? Well, ELIZA turned out to be too real for its (her?) own good.
ELIZA was created to act as a simulation of a psychotherapist by reacting to keywords in teletyped conversations with a human subject, offering follow-up questions pattern-matched to the previous statement. “My mother hates me,” for example, would be followed with “Who else in your family hates you?” Weizenbaum described ELIZA initially as a parody of this type of therapist, but he became appalled when test subjects began to genuinely open up to the machine, revealing emotionally sensitive information about themselves as if talking to a real person. Weizenbaum became so disturbed by the connections he saw people form with the automated therapist that he went on to write the infamous screed Computer Power and Human Reason: From Judgment to Calculation (1976), in which he argued against the anthropomorphization of machines with human characteristics.
ELIZA made history not only as the first successful attempt for a computer to mimic human-human interaction (as opposed to straight human-computer interaction), but also as a forerunner for early computer programming languages and the basis for the original computer role-playing games. So, gamers, you have a (robotic) therapist to thank for losing so many hours of your lives.
SHAKEY THE ROBOT: Problems? Solved
Named, it seems, by someone with a strong sense of humor, “Shakey” was built in 1966 by California’s celebrated Stanford Research Institute as the first A.I. able to handle both logical reasoning and the ability to move by itself. When put like that, Shakey sounds like the beginning of some Terminator-esque dystopian future, but the reality was far more mundane.
Shakey, whose short career saw upgrades and developments from ’66 through 1972 before being “retired from active duty,” had minimal mobility. He could move across floors (as long as they were flat enough for him to wheel around without falling over) and turn light switches on and off.
Remarkably, he could understand basic commands like “Push the block off the platform,” but also parse them into separate tasks, such as “Find the platform, find the block, push the block off the platform.” It might not sound like much compared with your average Cylon, but compared with anything that had come before, Shakey was the first small step—well, wheel trundle—into a brave new world. Instead of simply responding in a programmed way to a specific command, he was able to make plans, figuring out each small step in a command and stringing them together. This showed an understanding of the context and meaning of every command.
WATSON: Alex Trebek’s Worst Nightmare
Game show addicts and tech geeks the world over are familiar with Watson, the IBM supercomputer that thoroughly defeated two former champions on Jeopardy earlier this year. Not only did the computer—who, like the human competitors, wasn’t connected to any outside network during the contest—outperform the bones and red meat competition, but in doing so, it raised a million dollars for charity. Who says that computers have no compassion?
Watson’s week of Jeopardy! wins wasn’t just a ratings stunt, but the culmination of seven years’ worth of planning and development on behalf of the team that created him, including five years of test runs before his television debut. IBM developers weren’t the only ones who worked on the project; Carnegie Mellon University, University of Southern California’s Information Sciences Institute, and MIT were amongst the schools that offered expertise of faculty and students to help with Watson’s particular education, creating a leap forward in computers’ abilities to not only parse meaning from complex questions but also supply useful answers in so short a time that it can beat two well-known game show winners. If you think that sounds like a lot of work simply to build a better game show contestant, you’re right; IBM foresees a future for multiple next-generation Watsons as consultant machines in fields as varied as medicine, tech support, and business analytics.
From chess to Jeopardy!, parody psychotherapy to problem solving on the go, the field of artificial intelligence takes its leaps forward wherever it can find them, each one demonstrating that what might start as entertainment can end up changing the world, one robotic step at a time.
Spread the word!