The Girl With No Eyes

[originally published July 2003]

William Gibson may regret coining the term cyberspace in his 1984 novel Neuromancer. He received acclaim with the world of the Sprawl, which he created in the short story Johnny Mnemonic. But it was one well-tuned phrase,

jacked into a custom cyberspace deck that projected his disembodied consciousness into the consensual hallucination that was the matrix,

that helped win him the science-fiction triple crown: the Hugo, the Nebula, and the Philip K. Dick awards. Now he can’t get away from cyberspace, like an actor typecast by a too-successful performance in a role he may no longer love.

In this installment of Biblio Tech, we return to cyberpunk, which was very hot in the 1980s and retained considerable power throughout the 1990s. In the first decade of the 21st century, cyberpunk conjures much less, so this is an excellent time to give it a thoughtful look. Specifically, we’ll explore a particular theme of Gibson’s — namely, what distinguishes the human from the machine.

Human + Machine = ?

In computer science, the fascination with using technology for augmentation, particularly of the human intellect, is one of the oldest drivers in the field. Doug Engelbart introduced the term in the 1960s, but he credits Vannevar Bush’s seminal paper As We May Think, which The Atlantic Monthly published in 1945, for the inspiration. Englebart’s vision has proved incredibly influential, producing the mouse, the graphical user interface, hyperlinks, and online collaboration, among other things. Bush’s technological foresight may have been flawed in the details — our modern information systems are not based on microfilm, for example — but in the broadest sense, he got much of it right. He properly identified that information storage and retrieval would be one of the most important challenges facing those we now call knowledge workers.

Engelbart made his life’s work the solution of the augmentation problem — namely, how to make it easier for people to actually use mechanical aids to increase their capabilities. In the introduction to his 1962 report to the US Air Force on his research in this area, he wrote:

By “augmenting human intellect” we mean increasing the capability of a man to approach a complex problem situation, to gain comprehension to suit his particular needs, and to derive solutions to problems. Increased capability in this respect is taken to mean a mixture of the following: more-rapid comprehension, better comprehension, the possibility of gaining a useful degree of comprehension in a situation that previously was too complex, speedier solutions, better solutions, and the possibility of finding solutions to problems that before seemed insoluble. And by “complex situations” we include the professional problems of diplomats, executives, social scientists, life scientists, physical scientists, attorneys, designers — whether the problem situation exists for twenty minutes or twenty years. We do not speak of isolated clever tricks that help in particular situations. We refer to a way of life in an integrated domain where hunches, cut-and-try, intangibles, and the human “feel for a situation” usefully co-exist with powerful concepts, streamlined terminology and notation, sophisticated methods, and high-powered electronic aids.

In his work on augmentation, Engelbart invoked important examples to show that augmentation needn’t be simple amplification — as, for instance, a hammer does for our fist or a megaphone does for our voice — but rather, it could be abstraction and extension.

Threading through Gibson’s “Sprawl” stories (Johnny Mnemonic, Neuromancer, Count Zero, and Mona Lisa Overdrive) and later efforts (Virtual Light, Idoru, and All Tomorrow’s Parties) is an exploration of the boundaries and distinctions between humans and machines. Like a child with a box of mixed Lego kits, Gibson experiments with different combinations of pieces, creating monsters and angels and then exploring the potential relationships among them from different directions. From the concept of a person, he probes the implications of using technology for augmentation and the threshold that divides humans from machine. From the idea of the machine, he speculates on what added characteristics could turn an artificial intelligence (AI) into a human.

In Johnny Mnemonic, we encounter several boundary-testing experiments. In the story’s gigantic megalopolis, resulting from the fusion of the cities between Boston and Atlanta into the Boston Atlanta Metropolitan Area, or BAMA, human beings augmented with surgical implants are the norm. The lowest level of augmentation is the jack, the electro-optical connector that lets people connect their nervous systems directly to computers or vice versa.

The higher levels of augmentation we see in Gibson’s work seem to stem from an extrapolation of trends and visions in human prostheses. Although we take baby steps today toward mechanical ears and artificial eyes to help the deaf and the blind, consider a future in which we have perfected the ability to connect man-made devices to our nervous systems. How many people, given the opportunity, would choose to replace some imperfect pieces of their anatomies, not because they failed but just to achieve superior performance?

Some augmentation might not be visible or even operationally valuable, as in the case of Eddie Bax, the Johnny Mnemonic of Gibson’s title. Eddie has a memory device implanted in his head that lets him store data on his clients’ behalf, whether for safekeeping or for smuggling. The Lo Teks, whom we meet in Johnny Mnemonic, tend toward less functional augmentation — essentially, punk modifications like animal teeth, ears, and other changes made for shock value rather than performance enhancement. Gibson’s interest in tattooing and body piercing, also aspects of punk culture, are in evidence in various works, notably Virtual Light. It’s a small conceptual step from a grotesque tattoo to a dog’s ears grafted on a character’s head.

She seems to be staring …

Other augmentation in Gibson’s work is deliberately, even shockingly, visible and extravagantly useful to the augmented person. Molly Millions is one such. She’s invested a fortune in surgical implants to turn herself into a lethal fighting machine, a fortune that she earned by practicing several unsavory professions, including the oldest one. In the tip of each finger is a retractable knife blade, implanted by one of the best of Chiba City’s “black clinics.” Her nervous system is enhanced, rendering her perceptions and reactions lightning fast. Finally, and most strikingly, the prosthetics replacing her eyes combine vision, see-in-the-dark sensors, and computer interfaces, all covered by chrome covers that look at first glance like high-tech reflective sunglasses. Gibson exploits the shock value of this self-mutilation: Molly has superior eyesight and looks incredibly cool with her mirrored eye covers, but wow!

In 1995, Robert Longo made Johnny Mnemonic into movie starring Keanu Reeves. The movie wasn’t particularly successful, although it does have a very attractive star and several engaging elements. My personal beef with it is the rebalancing of the Molly / Eddie dynamic. In the short story Molly is tough and lethal whereas Eddie is a self-described “technical boy” whose one foray into crudeness flops until Molly rescues him. The Molly character that Gibson creates in the Sprawl novels has a lot of potential, and I earnestly hope that the rumored Neuromancer project doesn’t make the same mistake by submerging the killer queen again.

Gibson’s fascination with Molly is evidenced by the fact that unique among the characters he creates for the Sprawl stories, she spans all of them. She’s young and ambitious and serves as the love interest of several other characters in the early stories. In later ones, though, she’s old and cynical, but just as deadly. Why is Molly so important to cyberpunk? Certainly her sexuality is important to the success of Gibson’s early writing, but is that all? I think not. Her integration of technological, albeit not intellectual, augmentation is total and permanent. Her partner in Neuromancer is Case, the console cowboy. His augmentation is intermittent; he’s only augmented when he’s jacked into the matrix. Other times, he’s merely human.

One of the most fascinating experiments of Gibson’s work with Molly and Case comes when he outfits Molly with a sim/stim rig that transmits all her sensory inputs to Case. Suddenly the partnership has Case’s integration with the matrix and Molly’s integration with the physical world.

Machine + Augmentation = ?

In Neuromancer Gibson’s focus is on the quest by a machine, an AI, to augment itself. Throughout the novel, we encounter the efforts of one AI to merge with another, something that the Turing Police are systematically, though incompetently, constituted to prevent. Woven through this is a hard-boiled adventure yarn whose plot twists and confusions would do credit to Raymond Chandler or Dashiell Hammett.

Gibson raises some interesting questions. In Neuromancer, we encounter an AI with Swiss citizenship:

“It owns itself?”

“Swiss citizen, but T-A own the basic software and the mainframe.”

“That’s a good one,” the construct said. “Like, I own your brain and what you know, but your thoughts have Swiss citizenship. Sure. Lotsa luck, AI.”

This is the crux of the question. When a true AI actually comes to be, whether by accident or design, what rights should it have? Who will protect these rights? What will be its attitudes toward the human race? Gibson is not the first to ponder this topic, of course, as we discussed in the first Biblio Tech, but he does seem to have articulated and explored many more different aspects of the question in fictional scenarios.

Machine + Human = ?

In Count Zero we meet another augmented person in the form of Josef Virek. He’s a man whose body’s failure has been arrested but not stopped by the continuous addition of machinery. The novel hints that the augmentation’s primary purpose is preserving Virek’s life, but it is also clear that Virek has gained a certain level of multitasking and has lost some control over some of the manifestations of his persona in the process. This raises an interesting question: Is he still human? What does it mean to be human? Do we have to be a biological entity residing in a body? How much machinery can we add without sacrificing our humanity? Must these functions be provided biologically?

Eddie Flatline, whom we meet in Neuromancer, is a ROM construct — a recording of a dead console cowboy’s personality and memories. At one point, Flatline asks Case, a natural human, to destroy the ROM containing his personality, meaning the ROM construct has enough self-awareness to request death. This is a notion we encountered much earlier in Vernor Vinge’s True Names, when at the end, Erythrina records herself in a computer network’s data space. Explaining herself to her erstwhile but now uncertain ally, Mr. Slippery, she says,

When Bertrand Russell was very old, and probably as dotty as I am now, he talked of spreading his interests and attention out to the greater world and away from his own body, so that when the body died he would scarcely notice it, his whole consciousness would be so diluted through the outside world.

Lawyers have the term natural person to distinguish between corporations and people, because in a certain sense we have created corporations for the purpose of investing them with some of the rights and privileges of people. Perhaps we will be able to persuade an attorney with a theoretical bent to write about this for a future installment of Biblio Tech.

The ultimate reunion of the star-crossed lovers Bobby Newmark and Angela Mitchell in Mona Lisa Overdrive comes only after the deaths of their bodies and the transfer of their personalities into AIs destined to live in the Aleph’s context. In fact, it’s Virek’s quest to acquire the technology to permit that same transfer for himself that precipitates the entire sequence of events in Count Zero and Mona Lisa Overdrive, although Virek himself doesn’t survive the first episode.

As we’ve observed earlier, Gibson isn’t the first to have speculated on the use of AI as a framework for the preservation of the human (the soul?) after death, but he’s certainly the first to render it a casual assumption.

Love Not Human

Idoru‘s thesis is that a human and an AI fall in love and decide to marry. The novel spends its time and energy keeping us engaged in an attempt to grasp this point. The other characters in the story are engaged in various efforts to understand, thwart, or encourage the match.

Gibson is not the first to explore notions of emotional attachment between humans and AIs. Robert A. Heinlein established several close friendships between Mike and the humans most involved in setting up the Lunar revolution in his book, The Moon Is a Harsh Mistress. Mr. Slippery clearly maintains a personal loyalty to Erythrina even after her persona migrates permanently to cyberspace and she ceases to have a physical presence.

In these earlier stories, however, the authors maintained a clear distinction between the personalities living in the machines, whether they originated there or not, and natural humans. In Idoru, however, Gibson deliberately invokes aspects of love that we associate with bodies. Rei Toei, the artificial person, was originally constructed to be a performer. She is manifested as a holograph and appears as an attractive young woman. Rez, the human who wants to marry her, is a successful pop music performer with fan clubs on all continents, so his sudden obsession — his sudden crazy obsession — causes consternation among his friends, managers, and fans. His fascination with Rei Toei has an implicit carnal aspect that makes everyone squirm.

What precisely is marriage between a natural person and a virtual one? Gibson takes pains to make it clear that this is not the Platonic love between man and machine explored by Asimov, Heinlein, and others. This is the real thing. Unfortunately, Gibson walks to the brink but doesn’t jump. He leaves the consummation of the union unexamined at the end of Idoru, as is his right. But in the next installment of the story, All Tomorrow’s Parties, he cheats — when that scene opens, the two have parted company. Worse yet, by the end of that novel, he permanently eliminates the question by means of a deus-ex-machina maneuver that would be irritating if it weren’t such a sublime pun.

Why Cyberpunk?

What’s fascinating about Gibson’s writing is the focus on him as a literary stylist rather than as a speculator on the relationships between humans and their creations. Is this because the critics are largely littérateurs, primarily concerned with the world of words and uncomfortable with attempts to analyze the technological dimensions of Gibson’s work? Or is it because many of the ideas explored in his writing aren’t terribly new, as we’ve discussed in earlier articles?

Gibson’s success to date has been driven more by the punk than the cyber in his world. His artful creation of a jarring, dissonant dystopia is compelling; the technology is more of a veneer. Nonetheless, he has managed to touch and speculate on a collection of important questions that we as technologists should think about. Gradually, we will develop the ability to integrate machine and man; in fact, we’re doing it already with work in prosthetics and artificial intelligence. Because the process will be gradual, we are in danger of letting it happen unexamined. Each incremental step will benefit someone somewhere, and we will manage to avoid thinking about the systemic implications until suddenly we’re in an alien world that might well resemble one of Gibson’s nightmares.

That said, it’s important to recognize part of Gibson’s power as a writer is the power of the professional prestidigitator. His art is in misdirection, not magic. The worlds of Gibson’s writing are dystopic, with many foundations of our present world absent or disturbingly warped. Security comes from powerful allies, never from neutral institutions dedicated to maintaining the public good. Relationships that last are built on raw power, while balanced relationships are evanescent. This isn’t to say that comfortable homey things don’t exist in the Sprawl or in the Virtual Light world, but Gibson definitely makes sure we don’t see much of them.

Whenever we encounter children, as we do in several places, they are either street urchins living by their wits or sheltered flowers of the wealthy, as in the case of 13-year-old Kumiko Yanaka. We meet Kumiko in Mona Lisa Overdrive as she is being sent by private jet for safekeeping in London while her father, some sort of big shot in the Yakuza, sorts out some pending unpleasantness. Nothing about her life is what we would think of as normal. We see no school, we hear of no friends, but we do learn about her mother — albeit only her suicide — and Kumiko’s ambiguous feelings toward her father, whom she blames. Even when we encounter middle-class children, for instance the Tokyo Lo/Rez fanclubs in Idoru, we don’t see the prosaic day-to-day material of school and home that establishes context.


Developmental psychologists tell us that a child’s growth is characterized by an increasing ability to distinguish the self from others and from the world. Gibson’s writing, particularly in the Sprawl stories, explores breaking down that distinction between the self and the other. The console cowboy jacks into and merges with the matrix, being augmented and augmenting in turn. Sim/stim lets couch potatoes share the experiences of the stars, but it also lets Molly and Case achieve a new level of partnership. Wintermute seeks to merge with Neuromancer to create a new level of personality. Virek seeks to migrate his persona from his failing physical body to the immortal realm of the aleph. As quantum mechanics, via uncertainty, made hard little electrons into vague fuzzy presences, Gibson makes his people into fuzzy personas — not by making them vague and indistinct, but by blurring their boundaries. We keep coming back to the gist of his question: Just what is a person?

The people who occupy Gibson’s worlds are adrenaline junkies, criminals, mercenaries, and super celebrities, always living on the very edge. More than that, they are people who are completely foreign and, consequently, fatally fascinating, to the vast bulk of his readers. I’m indebted to Paul Brians of Washington State University, who notes that, “it is not surprising that he gained more of a following among academics than among the sort of people he depicted.”

One of the most fascinating speculations in both philosophy and computer science is over whether the human brain is a machine. If it is, then ultimately we can build a machine with equivalent complexity and capability and duplicate its every capacity, including creativity, imagination, vision, and boredom. If it is not, then some functional process in the brain, as yet not clearly demonstrated, must distinguish it from a computer. Nothing we know about the brain’s physical machinery so far suggests that it has any capability that can’t be duplicated with mechanisms. If so, what prevents us from being able to create an AI equivalent to a person? There conceivably might be some process in the brain that transcends mechanisms, some mystical facility that operates by means we don’t yet know or perhaps cannot ever understand. Or there might be some complexity threshold that we haven’t yet passed with our machines.

In All Tomorrow’s Parties, Gibson gives his personal answer to this question when the artificial person Rei Toei says to Rydell:

“This is human, I think,” she’d said when pressed. “This is the result of what you are, biochemically, being stressed in a particular way. This is wonderful. This is closed to me.”

Here we see Gibson’s failure as a theoretician. Nothing that the idoru claims underpins the distinction between AI and human is plausible to computer scientists and engineers who have considered the topic. There are no biochemical processes that cannot be modeled or simulated using computers. This damp squib leaves us with the unsettling feeling that Gibson has dropped the ball. It’s at times like these that you realize Gibson belongs to the literary world, not the concept-mad world of science fiction, unlike his brethren Asimov, Heinlein, and Vinge.

This installment of Biblio Tech has been dedicated to the work of one person, William Gibson. More than that, however, the articles in this department to date have all been building toward this examination of Gibson’s work. This is fitting; given the influence that his work has exerted on the field of science fiction and the entertainment he has given so many of us. I hope these articles inspire you to read some of the important works we’ve examined and, more importantly, to think about some of the issues discussed. As engineers, computer scientists, and general technologists we are among the best prepared to consider these topics and anticipate the implications of the technologies we are developing. We have an obligation to do so and to engage non-technical people in discussion.

Influential Works

Medium Author Title Year of Original Publication
Article Vannevar Bush As We May Think 1945
Book Philip K. Dick Do Androids Dream of Electric Sheep 1968
Novella Vernor Vinge True Names 1981
Short Story William Gibson Johnny Mnemonic 1981
Film Ridley Scott Blade Runner 1982
Book William Gibson Neuromancer 1984
Book William Gibson Count Zero 1986
Book William Gibson Mona Lisa Overdrive 1988
Book William Gibson Virtual Light 1993
Film Robert Longo Johnny Mnemonic 1995
Book William Gibson Idoru 1996
Film Andy and Larry Wachowski The Matrix 1999
Book William Gibson All Tomorrow’s Parties 1999

Read the original …

(This article appeared originally in IEEE Security & Privacy in the July/August 2003 issue. This is substantially the same text, with some minor formatting changes to take advantage of the power of the online presentation plus a few minor wordsmithing tweaks. And the table has the original publication dates for the listed books, not the editions in print in 2003 when the article was published.)

Here’s a PDF (article-04-final) of the original article, courtesy of IEEE Security & Privacy.

Hey, Robot!

[originally published May 2003]

What area of research, development, and commercial activity owes more of its existence to the arts than robotics does? The word itself comes from an early 20th-century play; less than a decade later, an important film introduced an enduring fantasy concept of what robots look like. Shortly after that, but still before much significant technical research or development occurred in the field, science-fiction writers developed complex theories of robot behavior in stories that are still in print today.

In this installment of Biblio Tech, we’ll look at some of the arts that have shaped our notions of robots. We will see the deep roots these stories have in far earlier concepts that have little to do with engineering but everything to do with the human race’s fascination with creation.

R.U.R. (Rossum’s Universal Robots)

In 1920, Karel Capek completed his play, R.U.R. (Rossum’s Universal Robots); its first production in 1921 brought Capek worldwide renown and introduced the word “robot” to the English language. Some argue that if he had survived the era of Nazi domination in Europe, he would have received the Nobel Prize for Literature.

Rather than mechanical constructions, Rossum’s robots were more biological and chemical in their fabrication. Nonetheless, they are definitely the ancestors of our modern industrial gadgets. Is the distinction between human beings and machines that humans work to live while machines exist to work? If that’s the case, then Rossum’s robots definitely existed, or were at least built, to work. They worked tirelessly and were tremendously more productive than mere humans, but they lacked emotions, creativity, and souls.

In the play’s first act, Helena Glory, the young daughter of “the President,” arrives by ship to the remote island where Rossum has developed the techniques for making robots. She is concerned with the oppression of robots worldwide and wants to foment a revolt among them — to inspire in them a passion for freedom.

What she finds on the island is a factory almost entirely staffed by robots, with a small team of men managing the operation. She is dismayed to discover that the robots are emotionless and unmovable: Rossum and son’s original engineering work produced a simplified physiology and nervous system that were incapable of pain or passion.

All is not lost, however. Dr. Gall, the head of the Experimental Department, is working to add a pain sense to the robots:

Helena: Why do you want to cause them pain?

Dr. Gall: For industrial reasons, Miss Glory. Sometimes a Robot does damage to himself because it doesn’t hurt him. He puts his hand into the machine, breaks his finger, smashes his head, it’s all the same to him. We must provide them with pain. That’s an automatic protection against damage.

In addition, there’s a mysterious disease called “Robot’s cramp” that the managers view as a fatal failure: “A flaw in the works that has to be removed.” Helena recognizes it as something else, though: “No, no, that’s the soul.”

In the remainder of the play, we watch the world’s economies devastated by cheap labor and see governments wage war with armies of robot soldiers. Finally, the robots revolt, ultimately exterminating their creators. The play ends with the emergence of a robot Adam and Eve and the cycle of life begins again.

Frankenstein, The Golem, and Metropolis

We see in Mary Shelley’s 1818 novel Frankenstein (unlike the flood of cinematic caricatures that sprang from it) a set of concepts similar to those in R.U.R. Behind Frankenstein, we see the even older legend of the Golem. In 16th-century Prague, the story goes, Rabbi Loew created a humanoid figure out of clay and brought it to life by marking it with a powerful magic word. He then commanded this creature to defend the Jews of the Prague ghetto against the torments of a contemporary despot. Ultimately, the creature began to show signs of rebellion. The ending of the legend has many different variations. In some versions, Rabbi Loew destroys the Golem; in others, the creature flees, never to be seen again.

Fritz Lang’s 1927 film Metropolis introduced the first cinematic robot, which managed to typecast the entire category for at least 50 years. Lang’s robot is the creation of a mad scientist, Rotwang, who is trying to create a surrogate for his lost love, Hel, to whom he has built an altar in his laboratory. She rejected him in favor of his rival, Joh Fredersen, and died giving birth to their son, Freder Fredersen. Joh Fredersen is the master of the city of Metropolis, an architectural and industrial vision of the early 20th century that we might barely recognize today. Metropolis is divided into two parts: a lower part inhabited by industrial workers who live underground and toil ceaselessly in the bowels of the machines that make Metropolis function, and an upper part peopled by a happy leisure class who spend their time at games and diversions. Near the beginning of the film, Freder ventures below ground, where his heart is moved by the plight of the workers and captivated by the beautiful Maria, a pure and gentle young woman whom he encounters preaching peaceful change. She promises a bridge for the gap between the workers (the Hands) and the managers (the Head). She calls this as-yet-unknown person the Mediator and identifies him as the Heart.

To undermine the workers’ movement, Joh has Rotwang give the robot Maria’s appearance. The robot then proceeds to rouse the workers to violence, which backfires when their children are threatened by floods unleashed by the destruction of some of the machines. Freder and the real Maria rescue the children, and the mob then burns the robot at the stake as Freder brokers a reconciliation between Joh and their leader.

The robot is referred to as the Machine-Man in the English intertitles before it is transformed into Maria’s sinister double. The double is everything that a thousand subsequent movie robots ever were: destructive, soulless, and ultimately evil. This movie is one of the most influential achievements of 20th-century filmmaking; you can see its influences in many subsequent cinematic masterpieces, as well as nearly every third-rate monster flick.

A common theme running through all these early stories is the classical Promethean notion that certain things are not meant for humans to control. Tampering with them trespasses on the domain of the divine and exposes the trespasser to severe punishment. Mary Shelley, in the preface to the 1831 edition of Frankenstein, wrote,

“Frightful must it be; for supremely frightful would be the effect of any human endeavor to mock the stupendous mechanism of the Creator of the world.”

Why is it that these stories — from the Golem to Frankenstein to Metropolis — always adopt classical models? Creating something that is alive or seems to be alive is portrayed always as trespassing on the perquisites of the divine, which is hubris and is certain to be punished by the gods. A simple explanation is that every storyteller tries to create a fiction that meshes with the real world — in this case, a real world in which intelligent robots are manifestly absent. To be complete, then, each story must end with a world without such things and a reason for their absence. You might ask, “But why aren’t there any man-made intelligences?” to which the answer would be, “Because there shouldn’t be, of course.”

In the 20th century, however, technological progress started to undermine the tyranny of “cannot.” Let’s look at the effect of that change on “should not.”

I, Robot

In 1939, a young man with a BS in chemistry from Columbia University wrote a story called Robbie about a little girl’s robot playmate. In a retrospective article about this and his other robot stories, entitled My Robots, Isaac Asimov said,

“In that case, what did I make my robots? I made them engineering devices. I made them tools. I made them machines to serve human ends. And I made them objects with built-in safety features. In other words, I set it up so that a robot could not kill his creator, and having outlawed that heavily overused plot, I was free to consider other, more rational consequences.”

In making them “to serve human ends,” Asimov didn’t innovate. However, in delving more deeply into their construction, particularly into their cognitive construction, he broke new ground.

Asimov went on to earn a PhD in biochemistry and work in academia teaching science, all the while writing science fiction throughout his long career. He brought to his writing tremendous conceptual power and a deep theoretical orientation. He was renowned as a prolific writer who could turn out a story or a book in a startlingly short time period, but this speed came at the expense of quality.

Much of Asimov’s early writing was not his best. The characters in the short stories that make up I, Robot are flat, the dialog wooden, and the best of the plots contrived. He did have exceptional moments in those early days when his writing soared — for example, in Nightfall — but in his youthful work this was the exception rather than the rule.

Nevertheless, the stories in I, Robot are important works, because in addition to repudiating the divine “You may not mock the stupendous mechanism of the Creator of the world” taboo, Asimov made a more fundamental contribution — namely, the Laws of Robotics:

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  2. A robot must obey orders given to it by human beings except where such orders would conflict with the First Law.
  3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

Suddenly, the issue was not about the sin of creating robots: it was about how to manage them appropriately. The importance of Asimov’s Laws of Robotics was not their precise formulation or wording, but that they existed at all. Engineered things, Asimov tells us, can be made subject to strict controls that aren’t applicable to humans. This is why constructing robot intelligence is not a sin, he says, any more than constructing anything else is a sin. Yes, we must pay attention to complicated details, but difficulty is not impossibility. Check with the Wright brothers, Sir Edmund Hillary, and one or two others if you doubt that fact.


With Bolo, Keith Laumer introduced the robot’s viewpoint. In this series of stories, begun in 1960 and largely complete by 1969, we encounter a series of robotic war machines — the evolutionary descendants of tanks. Laumer wasn’t the theoretician that Asimov was and the logic driving his thinking isn’t particularly transparent, but the concept is compelling.

In the Bolo stories, Laumer inserts sections of first-person monologue by the robot. This is a big step away from Lang’s notion of the robot as incomprehensibly alien — the Other. Instead, the robot thinks about its situation and reasons about the circumstances in which it finds itself. Laumer’s robots are invariably loyal to their human masters, although in Rogue Bolo, we encounter a robot with sufficient intellectual power to conduct a strategic campaign against adversaries that humans haven’t detected, despite direct orders from humans to desist. Implicit in this is Asimov’s assertion of the First Law’s precedence over the Second Law.

Star Wars

Released fifty years after Metropolis in 1977, Star Wars struck another small blow in the struggle to liberate robots from their earlier stereotypes as humanoid, ruthlessly competent, and evil. In this movie, we get a humanoid robot — C3PO — that is trivial and cowardly, though still part of the good guy crowd, in contrast to the lumpish and purely functional (but invariably competent and heroic) R2D2. C3PO is articulate whereas R2D2 is completely wordless, thus providing the ultimate cinematic example of the old saw that actions speak louder than words. Interestingly, in the recently released back-story, The Phantom Menace, we learn that the young Anakin Skywalker constructed C3PO. R2D2’s origin seems to be more prosaic, but there is some sort of justice in the fact that the weak C3PO was built by the person who turns out to be the penultimate bad guy. Perhaps C3PO’s weakness is a foreshadowing of Anakin’s own? It’s worth noting that by 1977, robots were so well established that this convergence of two separate themes — the fiction-inspired C3PO and the reality-inspired R2D2 — merits no more than a minuscule side plot in a science-fiction film.

Blade Runner

With Ridley Scott’s 1982 movie Blade Runner, loosely based on Philip K. Dick’s 1968 novel Do Androids Dream of Electric Sheep?, we return from mechanical humanoids to the biological creations Capek pioneered in R.U.R. Superficially, this is an exercise in which androids, called replicants and physically indistinguishable from humans, rebel against a social order that treats them brutally. They are banned from Earth — a formula that Asimov used to great effect in his robot stories — and have artificially limited lifespans. Their superhuman physical and mental capabilities are key to detecting them when they run and hide. We learn in Blade Runner that they fear death and so run to seek freedom and an unimpeded lifespan.

The understatement built into Blade Runner is overwhelming. Are replicants human? Their bodies are biological and they look like people, so it’s too easy to grant them souls by dismissing their creation as some perversion of cloning. Hannibal Chew, the engineer who boasts, “I design your eyes,” to two replicants right before they kill him, refutes this: if he’d just cloned their eyes, how could they have superior eyesight? And Harrison Ford’s character, Deckard, is a paradox: How can a human, every other instance of whom is manifestly inferior to replicants in physical and intellectual capabilities, manage — unaided — to defeat an entire team of replicants, one after another?

Moreover, Blade Runner re-poses the same question that R.U.R. asked: can you create an entity with intellectual capabilities and not give it a soul? Deckard speculates on this at the end of the movie while reflecting on a replicant’s decision not to kill him when he’d won the final fight:

“I don’t know why he saved my life. Maybe in those last moments he loved life more than he ever had before. Not just his life, anybody’s life, my life. All he’d wanted were the same answers the rest of us want. Where did I come from? Where am I going? How long have I got? All I could do was sit there and watch him die.”

Meanwhile, in the real world …

The golden age of robotics research came to an end sometime in the mid 1980s when a pair of economists observed that the sweet spot for flexible automation was in an area in which US industry took no interest. It turned out that robots are cost-effective for production runs roughly between 1,000 and 10,000 units. US manufacturing tends to have its sweet spots below 1,000 (airliners, electric generators, and supercomputers) and above 100,000 (jelly beans and automobiles). Japan’s manufacturing industry has historically focused its attention on 1,000 to 10,000 unit runs, giving it a tremendous ability to respond to market dynamics with revised products and simultaneously making robotics a far more economically attractive proposition. The result of this economic insight was a dramatic drop in research funding for robotics worldwide. Nonetheless, the field has made substantial technical progress in the past 20 years, albeit largely out of the public eye. Interestingly, there hasn’t been the same attention to robotics in the science fiction community, at least not in the works that have gotten attention from the broadest community of readers.

Is this parallel drop-off in the world of fictional robots because Asimov and Laumer said everything there is to say about robots? Is it that people have recognized the absurdity of humanoid robots and have transformed the debate into one about the broader topic of artificial intelligence, as we considered in the first Biblio Tech article? Or are we just bored with the topic? I’m not sure. I prefer to think that we’re just waiting for some powerful new talent to turn our thinking on its head again with a brilliant new insight.

SIDEBAR: What is a robot?

There is no real consensus on precisely what a robot is. Rather than trying to define one, let’s instead try to identify the characteristics of things that we might call robots. The most appealing description is
that a robot is a system with mechanical components intended to achieve physical action; it also has sensory feedback and a sophisticated and flexible control system that links its sensing to action.

To see if this characterization works, let’s see if it correctly distinguishes between our ideas of robots and nonrobots. The system must be intended to produce mechanical action, so a computer video game is out. The system must use sensory feedback to control motion, so printers are out. So far, there is little to distinguish a robot from a classical control system.

A numerically controlled machine tool is a robot, but just barely. A modern car’s antilock braking system could qualify, although there’s something unsatisfying in it doing so. An airplane’s autopilot certainly qualifies as a robot, particularly the advanced autopilots that can receive a list of waypoints and then navigate themselves from liftoff to approach via GPS. A washing machine that can sense the amount and temperature of water in its tub and act accordingly is probably a robot, albeit not a particularly interesting one. Oddly enough, many of the pick-and-place industrial robot systems in factories in Japan and elsewhere around the world fail this test, because they lack a sensory capability.

Today, robots are almost commonplace. We see them in numerous prosaic roles in factories, but we also see them competing in what can best be called a new form of demolition derby. Students around the world work to build robots to compete in robot soccer; a researcher at Bell Labs built one to play ping-pong a few years back. Numerous toys on the market incorporate various aspects of robotic technology.

Influential Works

Medium Author Title Year of Original Publication
Book Mary Shelley Frankenstein 1818
Play Karel Capek R.U.R. (Rossum’s Universal Robots) 1920
Film Paul Wegner, Carl Boese Der Golem (in German) 1920
Film Fritz Lang Metropolis 1927
Book Isaac Asimov I, Robot Short stories: 1940-1950; collection: 1950
Book Isaac Asimov Nightfall 1941
Book Keith Laumer Bolo Short stories: 1960-1976; various collections
Book Philip K. Dick Do Androids Dream of Electric Sheep? 1968
Film George Lucas Star Wars 1977
Film Ridley Scott Blade Runner 1982
Book Michael Chabon The Amazing Adventures of Kavalier and Clay 2000

Read the original …

(This article appeared originally in IEEE Security & Privacy in the May/June 2003 issue. This is substantially the same text, with some minor formatting changes to take advantage of the power of the online presentation plus a few minor wordsmithing tweaks. And the table has the original publication dates for the listed books, not the editions in print in 2003 when the article was published.)

Here’s a PDF (article-03-final) of the original article, courtesy of IEEE Security & Privacy.