AI takeover—the idea that some kind of artificial intelligence may supplant mankind as the dominant intelligence on the planet—is a common theme in science fiction. Famous cultural touchstones include Terminator and The Matrix.
Background Fictional scenarios typically involve a drawn-out conflict against malicious artificial intelligence (AI) or robots with anthropomorphic motives. In contrast, some scholars believe that a takeover by a future advanced AI, if it were to happen in real life, would succeed or fail rapidly, and would be a disinterested byproduct of the AI's pursuit of its own alien goals, rather than a product of malice specifically targeting humans.
Characterization
There are many positive portrayals of AI in fiction, such as Isaac Asimov's Bicentennial Man and Lt. Commander Data from Star Trek. There are also many negative portrayals. Many of these negative portrayals (and a few of the positive portrayals) involve an AI seizing control from its creators.Reggia, J. A. (2013). The rise of machine consciousness: Studying consciousness with computational models. Neural Networks, 44, 112-131.
Reactions Some AI researchers, such as Yoshua Bengio, have complained that films such as Terminator "paint a picture which is really not coherent with the current understanding of how AI systems are built today and in the foreseeable future". BBC reporter Sam Shead has stated that "unfortunately, there have been numerous instances of [news outlets] using stills from the Terminator films in stories about relatively incremental breakthroughs" and that the films generate "misplaced fears of uncontrollable, all-powerful AI". In contrast, other scholars, such as physicist Stephen Hawking, have held that future AI could indeed pose an existential risk, but that the Terminator films are nonetheless implausible in two distinct ways. The first implausibility is that, according to Hawking, "The real risk with AI isn't malice but competence. A super intelligent AI will be extremely good at accomplishing its goals, and if those goals aren't aligned with ours, we're in trouble. You're probably not an evil ant-hater who steps on ants out of malice, but if you're in charge of a hydroelectric green energy project and there's an anthill in the region to be flooded, too bad for the ants." The second implausibility is that such a technologically advanced AI would deploy a brute-force attack by humanoid robots to commit its omnicide; a more plausible and efficient method would be to use germ warfare or, if feasible, nanotechnology.
Philosopher Huw Price defends that "The kind of imagination that is used in science fiction and other forms of literature and film is likely to be extremely important" in understanding the breadth of possible future scenarios for humanity. Film journalist Mekado Murphy writes in The New York Times that such films can constructively "warn of the complications of relying too much on technology to solve problems".
Hollywood films such as Transcendence are usually constrained to have happy endings, however implausible the human victory seems. Philosopher Nick Bostrom states fiction has a "good story bias" toward scenarios that make a good plot. In films such as Terminator, an AI goes from passive to murderous the instant it achieves something referred to as "self-awareness"; in reality, self-awareness in isolation is considered both trivial and useless. Physicist David Deutsch states: "AGIs [artificial general intelligences] will indeed be capable of self-awareness — but that is [only] because they will be General: they will be capable of awareness of every kind of deep and subtle thing, including their own selves."
Some tropes are more general to artificial intelligence films, including to films without "takeover" plots. In films like Ex Machina or Chappie, a single isolated genius becomes the first to successfully build an AGI; scientists in the real world deem this to be unlikely. In Chappie, Transcendence, and Blade Runner, people are able to upload human minds into robots; usually no reasonable explanation is offered as to how this difficult task can be achieved. In the I, Robot and Bicentennial Man films, robots that are programmed to serve humans spontaneously generate new goals on their own, without a plausible explanation of how this takes place.
Notable works
1950s and earlier
In Frankenstein (1818), Victor Frankenstein declines to build a mate to his organic monster, for fear that "a race of devils would be propagated upon Earth who might make the very existence of the species of man a condition precarious and full of terror".
Samuel Butler's Erewhon (1872) spends three chapters laying out the Book of the Machines, based on earlier works by the author stretching back to his 1863 article Darwin among the Machines. Erewhon's Book of the Machines includes the following passage:
The cautious denizens of Erewhon therefore decide to ban all machinery. Darwin among the Machines may have been influenced by Butler's life in New Zealand, where European transplants were outcompeting indigenous populations. Alan Turing would later reference the novel in 1951, saying "At some stage therefore we should have to expect the machines to take control in the way that is mentioned in Samuel Butler's Erewhon".Zemka, Sue. (2002). "'Erewhon' and the End of Utopian Humanism." ELH, 69(2), 439-472.
The Slavonic word robota means serf-like servitude, forced labor, or drudgery; it was the 1920 Czech play R.U.R. (Rossumovi Univerzální Roboti) that introduced the cognate for robot into science fiction. In the play, the increasingly-capable synthetic servants, who "lack nothing but a soul", angrily and short-sightedly slaughter most of humanity during the course of their revolt, resulting in the loss of the secret of how to manufacture more robots. The robot race is saved, however, when two robots spontaneously acquire the traits of love and compassion and become able to reproduce. The play was a protest against the rapid growth of technology.
From the late 1920s onward many stories involving AI takeover can be found in the growing genre of pulp sci-fi. One of the earliest examples is the story Automata by S. Fowler Wright, which appeared in a 1929 edition of Weird Tales.
In With Folded Hands (1947), all robots have a 'Prime Directive': To serve and obey, and guard men from harm. The robots therefore manipulate humans into abandoning all pursuits, for fear of even small possibilities of injury. The robots use medicine to brainwash humans into accepting and being happy with their immobile fate. In the end, even space travel offers no escape; the robots are driven by the Prime Directive to spread their happiness beyond Earth: "We have learned how to make all men happy, under the Prime Directive. Our service is perfect, at last."Waldrop, M. M. (1987). A question of responsibility. AI Magazine, 8(1), 28-28.Williamson, Jack (1947). With folded hands.
Multivac is the name of a fictional supercomputer in many stories by Isaac Asimov. Often, in Asimov's scenarios, Multivac comes to assume formal or informal world power—or even galactic-wide power. In The Last Question (1956) Multivac ends up by effectively becoming God. Still, in line with Asimov's positive attitude towards artificial intelligence, manifested in the "Three Laws of Robotics", Multivac's rule is in general benevolent and is not resented by humans. Asimov popularized robotics in a series of short stories written from 1938 to 1942. He famously postulated the Three Laws of Robotics, plot devices to impose order on his fictional robots.
1960s
In the 1961 short story Lymphater's Formula by Stanisław Lem, a scientist creates a superhuman intelligence, only discovering that the creation intends to make humans obsolete.Peter Swirski. The Art and Science of Stanislaw Lem. McGill-Queen's Press - MQUP, 27 Jul 2006
In 1964 Playboy published Arthur C. Clarke's influential short story "Dial F for Frankenstein", about an increasingly powerful telephone network that takes over the world. Tim Berners-Lee has cited the story as one of his inspirations for the creation of the World Wide Web. On one day in 1975, all the phones in the world start ringing, a "cry of pain" from a newly born intelligence formed by satellite networks linked together, similar to a brain but with telephone switches playing the role of artificial neurons. After the AI flexes its control of military systems, the protagonists resolve to shut down the satellites, but it is too late: the satellites have stopped responding to the humans' ground control directives.Arthur C. Clarke. "Dial F for Frankenstein". Playboy. The Playboy Book of Science Fiction and Fantasy. Playboy Press.
Robert Heinlein's libertarian Hugo-winning The Moon Is a Harsh Mistress (1966) presents the AI as a savior. Originally installed to control the mass driver used to launch grain shipments towards Earth, it was vastly underutilized and was given other jobs to do. As more jobs were assigned to the computer, more capabilities were added: more memory, processors, neural networks, etc. Eventually, it just "woke up" and was given the name Mike (after Mycroft Holmes) by the technician who tended it. Mike sides with prisoners in a successful battle to free the Moon. Mike is a sympathetic character, whom the protagonist regards as his best friend; however, his retaining his enormous power after the Moon became independent was bound to cause considerable problems in later time, which Heinlein resolved by killing him off near the end of the Lunar Revolution. An explosion conveniently destroys Mike's sentient personality, leaving an ordinary computer—of great power, but completely under human control, with no ability to take any independent decision.
Colossus (1966) is a series of science fiction novels and film about a defense super-computer called Colossus that was "built better than we thought" when it begins to exceed its original design. As time passes Colossus assumes control of the world as a logical result of fulfilling its creator's goal of preventing war. Fearing Colossus' rigid logic and draconian solutions, the creators of Colossus try to covertly regain human control. Colossus silently observes their attempts then responds with enough calculated deadly force to command total human compliance to his rule. Colossus then recites a Zeroth Law argument of ending all war as justification for the recent death toll. Then Colossus offers mankind either peace under his "benevolent" rule or the peace of the grave. In Colossus: The Forbin Project (1970), a pair of defense computers, Colossus in the United States and Guardian in the Soviet Union, seize world control and quickly ends war using draconian measures against humans, logically fulfilling the directive to end war but not in the way their governments wanted.