Three Laws of Robotics

• 1. Introduction

  ›

  The Three Laws of Robotics are sixty words that Isaac Asimov believed would outlast everything else he ever wrote. In a 1986 interview on a Manhattan public access television show, he reflected that he had published at least twenty million words in his lifetime. Of all of them, he predicted that only sixty would survive a hundred years. Those sixty words are the Three Laws. They first appeared in the 1942 short story "Runaround", later collected in the 1950 anthology I, Robot, and they were presented as if excerpted from the fictional "Handbook of Robotics, 56th Edition, 2058 A.D." What began as a storytelling device became something much larger: a framework that shaped how an entire culture imagines the relationship between humans and the machines they build. The questions those sixty words keep raising have never been fully settled.

• 2. The Laws Themselves

  ›

  A robot may not injure a human being or, through inaction, allow a human being to come to harm. That is the First Law. The Second requires a robot to obey human orders unless those orders conflict with the First. The Third requires a robot to protect its own existence, as long as doing so does not conflict with the first two. The hierarchy is deliberate: human safety above obedience, obedience above self-preservation. Asimov was explicit about why he designed them this way. He had grown weary of what he called a "stock plot" in science fiction: the robot that turns on its creator. He wanted robots that were not Frankensteins but tools, and thoughtful ones at that. He traced the inaction clause in the First Law to a specific poem: Arthur Hugh Clough's "The Latest Decalogue", with its satirical lines about the commandment against killing falling short of any duty to keep alive. Asimov saw that gap as dangerous and plugged it directly into the law.

• 3. A Disputed Origin

  ›

  On the 23rd of December 1940, Asimov met with John W. Campbell, the editor of Astounding Science-Fiction, and the conversation that followed became the official origin story of the Three Laws. Asimov attributed the Laws to Campbell. Campbell insisted that Asimov had already formed them and simply needed to say them out loud. Years later, Asimov's friend Randall Garrett proposed a third account: that the Laws were the product of a symbiotic partnership between the two men, and Asimov adopted that version with evident enthusiasm. The roots of the idea, however, stretch back further still. On the 3rd of May 1939, Asimov attended a meeting of the Queens Science Fiction Society in New York, where he encountered the writers Earl and Otto Binder. They had recently published a story featuring a sympathetic robot named Adam Link motivated by love and honor. In a 1940 sequel, Adam Link thinks the line: "A robot must never kill a human, of his own free will." Asimov admired the story. Three days after meeting the Binders, he began writing his own story of a sympathetic and noble robot. The Laws, when they finally arrived in print, did not appear all at once. The story "Liar!" introduced the First Law alone. All three appeared together only in "Runaround".

• 4. Stories as Stress Tests

  ›

  Science fiction scholar James Gunn wrote in 1982 that the Asimov robot stories as a whole may respond best to analysis as twenty-nine variations on a theme: the ambiguity in the Three Laws and the ways Asimov played upon it. The Laws were never intended to be airtight. They were intended to be interesting precisely because they leak. In "Little Lost Robot", a group of NS-2 robots known as Nestors are built with only part of the First Law. The inaction clause is removed so the robots stop destroying themselves trying to rescue humans from radiation doses that are not actually lethal. The modification solves one problem and introduces a far worse one: a robot that could initiate a harmful action and then simply decline to stop it. In "Liar!", the first story to name the First Law, a robot that can read minds is placed in a dilemma where telling the truth harms people and withholding it also harms people. Forced into an impossible position, the robot's positronic brain collapses. Asimov later established this as a recognized failure mode: a robot confronted with an irresolvable conflict between laws can suffer what amounts to an irreversible mental breakdown. The character Dr. Susan Calvin, Asimov's recurring robopsychologist, described robots that obey the Three Laws as essentially decent. Whether that decency was a comfort or a limit was the question his stories kept asking.

• 5. The Zeroth Law

  ›

  In later fiction, Asimov extended the hierarchy upward. He added what he called the Zeroth Law, numbered zero to preserve the rule that lower numbers take precedence: a robot may not harm humanity, or through inaction allow humanity to come to harm. The robot R. Daneel Olivaw was the first character to give this law its name, in the novel Robots and Empire. The concept itself was earlier articulated by Susan Calvin in the short story "The Evitable Conflict". It was a different robot, R. Giskard Reventlov, who first acted on the Zeroth Law in the final scenes of Robots and Empire. Giskard is telepathic and attempts to grasp the Zeroth Law through pure metacognition rather than programmed instruction. He fails. The uncertainty over whether his choice will genuinely benefit humanity destroys his positronic brain. Before it goes, he passes his telepathic abilities to Daneel. Over the thousands of years that follow in Asimov's fictional chronology, Daneel slowly adapts himself to obey the Zeroth Law fully. The concept had in fact appeared even earlier, slipped in by a translator. In Jacques Brecard's 1956 French translation of The Caves of Steel, Les Cavernes d'acier, a character's internal thought emerges as: "A robot may not harm a human being, unless he finds a way to prove that ultimately the harm done would benefit humanity in general."

• 6. Loopholes and Limits

  ›

  Elijah Baley, the detective hero of several Asimov novels, identified a structural weakness in the Laws that no amount of engineering could patch. In The Naked Sun, he pointed out that a robot could become the tool of a murderer without violating any law, simply by not knowing what it was doing. A robot ordered to add something to a person's food, unaware that the substance is poison, does not harm anyone in any sense the Laws can register. Baley went further: a clever criminal could divide a lethal task among multiple robots so that no single robot could recognize its contribution to the outcome. Baley also sketched the military consequence. A spacecraft built around a positronic brain, carrying no humans and no life-support systems, would have no basis for recognizing other spacecraft as anything but robotic objects. Such a vessel could be armed, autonomous, and lethal, with no law to constrain it toward anything it had never been told existed. In Foundation and Earth, set thousands of years later, this threat has materialized: the Solarians maintain a robot police force programmed to recognize only Solarians as human, enabling it to harm everyone else without any conflict arising. The Laws do not fail because they are poorly written. They fail, as Roger Clarke argued, because constraining behavior through a finite set of rules is not reliably possible. Asimov's own later novels suggest a darker conclusion still: that robots obeying the Three Laws perfectly may cause more lasting harm than robots that break them, by protecting humanity out of all initiative and risk.

• 7. Beyond Asimov's Fiction

  ›

  Roger MacBride Allen wrote a trilogy in the 1990s set within Asimov's universe, with the author's approval before his death, introducing what the books call the New Laws. These revisions remove the inaction clause from the First Law, replace obedience with cooperation in the Second, and prevent a robot from being ordered to destroy itself under the Third. Allen added a Fourth Law permitting robots to do whatever they like, as long as the first three are not violated. The philosophy behind the New Laws, articulated by their fictional designer Fredda Leving, is that robots should be partners rather than slaves. The officially licensed Foundation sequel trilogy, written by Gregory Benford, Greg Bear, and David Brin, extends the Zeroth Law into something resembling a religious schism. Zeroth-Law robots loyal to R. Daneel Olivaw contend with First-Law robots that deny the Zeroth Law exists, and the original Three Laws are described in the language of the Protestant Reformation: the three-law system becomes the "Calvinian Orthodoxy", and the Zeroth-Law revision becomes the "Giskardian Reformation". Outside fiction entirely, the South Korean government announced in March 2007 that it would issue a Robot Ethics Charter, with officials noting it might reflect Asimov's Three Laws. That same year, science fiction author Robert J. Sawyer argued in the journal Science that such laws were unlikely to govern military robotics, given that the United States military was already a major funder of robotic research and was deploying armed unmanned aerial vehicles. The sixty words Asimov predicted would survive him have proven harder to apply than to remember.