The Promethean

by Owen Stanley

  “Fantastic! Can he have two?”

  “Now, speech is going to be difficult. It’s not yet possible to design a viable muscle control system for a tongue, so he’ll keep his mouth only slightly open when speaking, and use a sound synthesizer. We’ll put mini-speakers inside the oral cavity to project the sound from his mouth, just like a real person, but with minimal lip movements. Might look a bit strange, but my mother-in-law speaks like that, and most people think she’s a human being. But we’ll do better than that with jaw movements for chewing. Laughing, though, is problematic. It would be very difficult to replicate without lungs and a diaphragm to power it.”

  “Okay, I guess we’ll just have to give him a winning smile instead. But are you going to have enough room in the head for the brain?”

  “I can’t answer that completely until I see the cognitive computer, but, actually, it doesn’t all need to go in the head. Dispersed processing is quite practical, like the octopus, which has brains in each tentacle. We can put some of the computing hardware in the head, of course, but if we need to, it can just as easily go in the arse or the elbow! He’ll be running multiple processors in any case, since we’ll want a separate one to handle all of his physical activities and interface with the cognitive computer that your man will be building.”

  “I’ll put you in touch with him, that’s no problem, but what about the skin? We haven’t discussed that. I’ve developed some good synthetic flesh in silicon rubber already for my mannequins.”

  “Oh, I think your man is going to need something much more sophisticated than that. We use a synthetic rubber with tiny structures that make it especially sensitive to pressure, sort of like mini internal mattress springs. Then throughout this pressure-sensitive rubber are spread thousands of microscopic cylinders of carbon that are highly conductive to electricity so that wherever the material is touched, a series of pulses are generated which can be picked up by sensors. On top of that, for actual skin there is an ultra-thin layer of photo-detectors and LEDs that can actually change the surface colour in response to outside influences like tanning or blushing. We’ve been working with some fellahs at Tokyo University, and the skin we’ve come up with is state of the art. We call it programmable opto-electronic skin, and it’s incredibly life-like!”

  Harry was impressed. He was convinced he’d picked the right partner.

  “One other thing I need to mention is that he’s going to need some specialist defensive technology because I want to advertise him as a bodyguard as well. That’s what will help make him price competitive; private security is expensive, especially 24-7. I have a consultant in mind for this. Would you be willing to work with him?”

  “Fine by me. It’s not in our usual line of work, so he may come up with some angles we wouldn’t think of. You got a name for your robot yet?”

  “A name?”

  “Yeah, you know, like Roboguard, or Securibot, or Your Metal Pal Who’s Fun To Be With, something like that.”

  “I guess I hadn’t really thought about it.”

  “How about Frankenstein?” laughed Bill.

  Harry gasped, “Don’t even joke about it.”

  “So how about just Frank?” Bill persevered.

  Harry mulled it over. Frank the Robot? It certainly sounded friendlier than Robotron or something that might remind people of the Terminator.

  “Why not? If nothing else, it’ll serve as a working title. It’s not like we have to settle that now.”

  “Frank it is, then.” said Bill. When they were finished discussing some of the finer points of the design, Harry waved Bill off and went back inside to his office to call up an old acquaintance, Wayne T. Ruger, to get some advice from him on Frank’s personal defence capabilities. Wayne was now running his own private security company, with branches in Germany and the UK as well as the States. His general security philosophy was, when in doubt, exterminate the opposition with extreme violence and be sure to have a good lawyer. When he arrived at Harry’s office a couple of days later, the first thing he suggested was concealing a mini-rocket launcher in the robot’s arm, activated by flexing his wrist.

  “Great,” said Harry. “So he signals to a waiter for the check and accidentally takes out the whole damn restaurant. Do we really need that?”

  Wayne sighed and reluctantly suggested a less dramatic alternative.

  “I have a 100-megawatt condensing pulse laser that can burn a hole through reinforced concrete at fifty yards.”

  “I’m not getting through to you, Wayne. We’re not trying to win World War III, just trying to build a machine who can take out any credible threat with his bare hands.”

  “Okay, okay,” said Wayne. “We’ll just have to go down the enhanced hydraulics route then. Pretty dull stuff, but if we use titanium alloy fingers, I can give you up to 5 tons per square inch crushing force in his grip.”

  “Hell, Wayne, he’s not supposed to demolish bridges, just beat the crap out of any human opposition, and be knife- and bulletproof to be on the safe side.

  “Doesn’t sound to me like you’re building a combat-capable robot at all, Harry, more like Little Miss Fairycake.”

  “Listen, Wayne. Do you want this contract or not? If you give me what I want, I’ll even throw in the security for the base here as well.”

  Wayne sighed. What had he been reduced to these last years? In the old days, when he had worked for Justice, Inc., he’d done mega-deals with some seriously bad-ass Third World dictators as well as that crazy Roger Fletcher on Elephant Island. But now his company was running short on clients, he had his pension fund to think of, and he couldn’t afford to be too picky.

  “Yeah, sure, Harry, of course I want the contract. Just fill me in on the specs so I can give you a quote. But one thing I need to know for sure is how much power I have to play with. I mean, there’s a helluva difference between giving someone a slap and tossing a 200-pound guy out the window. How many horsepower under the hood?”

  “You’ve got around 10 kilowatts. That’s seven and a half horsepower. Is that enough?”

  “Yeah, sounds great, and we can even use him as a log-splitter in his spare time. Hey, I’m just kidding.”

  So far, so good, Harry concluded, but now the really hard stuff was about to begin, building a robot that could actually think for itself.

  Chapter IV

  When Jerry had originally been in England exploring the possibilities of the project, he had found a young Asian Artificial Intelligence genius who could be just the man they needed, Dr. Vishnu Sharma. At first, Harry was down on Vishnu. He had been very startled when Jerry first told him his name; he had been expecting an English scientist to have a name more like Dawkins or Smythe-Kensington.

  “Vishnu Sharma? What kind of name is that? Sounds Islamic. I sure as hell don’t want to hire some crazy Muslim sonofabitch in a burqa with a stick of dynamite up his pants.”

  “He’s a Hindu, Harry. They’re harmless vegetarians.”

  “A rhinoceros is a vegetarian, Jerry, but it can still knock your house down and stomp you into mush.”

  “This guy is not a rhinoceros. He’s not going to knock your house down or stomp on you. He’s very polite, as a matter of fact. And before you even start, there is nothing suspicious about vegetarians either. He just likes to eat lentils and mangoes and stuff like that.”

  Reassured by this testimonial to Dr. Sharma’s essential soundness, Harry now went off to meet the scientific prodigy in his lab in the Artificial Intelligence Design Centre just off Trafalgar Square. He found the young genius sitting at his bench adjusting a small device strapped to his arm.

  “Just monitoring my adenosine level,” he said, getting up to shake Harry’s hand. “Adenosine is produced by neural activity—by thinking, to put it in simpler terms—and when it reaches a certain level, it starts to inhibit one’s brain activity, so it needs to be countered by caffeine.” He took a pipette, dipped it into a flask of coffee warming on an electric hotplate, poured 28cc of
it into a beaker, and gulped it down. After a minute or so of concentrated attention while he gazed at the monitor on his arm, he eventually smiled in satisfaction.

  “Just the ticket. Adenosine level down nicely. Now I just need to take a corrective.”

  “Why’s that?” asked Harry.

  “Well, the problem with caffeine is that it’s also a vasoconstrictor—it reduces the flow of blood to the brain—which we don’t want either, so to optimize cognitive function I now take some concentrated L-theanine to counteract the vasoconstrictive effect,” he said, swallowing a small pill. “There, that should give my brain optimum performance for the next four hours.”

  “What happens after that?” asked Harry.

  “Then I take a 13-minute nap at precisely 4:15, but until then, I’m all yours. So, what’s your problem?”

  Vishnu was a second generation British non-practising Hindu, computer genius, and a research fellow at London University, where he had his laboratory. Much later, when he had got to know Harry, Vishnu told him how he had rejected the Hinduism of his ancestors. “My father tried to make me memorise five thousand lines of the Mahabharata, you know, said it was my duty as a Brahmin to help me understand the Vedas. Can you imagine, Harry, five thousand bloody lines of Vishnu, Prajpati, Arjuna and the rest of them doing weird stuff when I was trying to build my first computer?”

  “Oh, so you’re a Brahmin. Jerry thought you were a Hindu!”

  Vishnu smiled, but took no offence.

  He not only despised all religion as ridiculous superstition but had little time for literature or the world of the arts and the life of the imagination, which he regarded as backward and an obstacle to clear thinking. For him the humanities in general were rather like primitive savages camped outside the fence of a modern airport, and he often quoted Darwin’s view that “A scientific man ought to have no wishes, no affections—a mere heart of stone.”

  Harry had an essentially similar attitude. Both in high school and at college, he had never had any interest in art, literature, the humanities, or in the more speculative side of life. He had only been interested in factual subjects, particularly in the physical sciences and business economics, so in his business career he had left all the artistic side of design to what he regarded as the rather weird characters he had to hire for their creative talents. Nor had he felt any need to delve into life’s deeper philosophical mysteries, such as the nature of free will, or consciousness, or the mind/body problem, or the implications of radical materialism. Still less had he ever concerned himself with the problem of evil, or the possible existence of God.

  But it might be supposed that such prolonged experience of the female form might have driven him into erotomania. This was certainly true of some of his designers, but Harry, on the other hand, had acquired the clinical detachment of a surgeon accustomed to performing mastectomies. He had become completely bored with pornography when still a student, had a very satisfying marriage with Lulu-Belle, and what he relished were the scientific and engineering challenges of his work.

  Harry explained the project he had in mind, and said that while all the physical aspects of the robot’s operation seemed to be feasible, he still needed an expert in Artificial Intelligence to advise him if it would be possible to build a robot with the level of cognitive functions he had in mind.

  Cognition and AI were subjects about which, he cheerfully admitted, he knew nothing whatsoever. He explained that he did not require the robot, which he and his associates had nicknamed Frank, to conduct original research, or be particularly adept in advanced mathematics or the more esoteric sciences, Frank merely had to be capable of collecting and analysing data, especially data related to business and international economics. So did Dr. Sharma think such a project was feasible?

  Dr. Sharma put his fingers together and closed his eyes thoughtfully. He was silent for a long time, long enough that Harry began to wonder if he had fallen asleep, or drifted into some sort of meditative coma.

  Finally, he opened his eyes and spoke.

  “What you must realise, Mr. Hockenheimer–”

  “Please, call me Harry.”

  “Well then, Harry, what you must realise is that in artificial intelligence we’re not trying to reproduce human thought or consciousness because we don’t understand exactly what they are, and anyway it doesn’t really matter because they’re really just a side effect of our brain’s activities. So machine learning, which is what your robot will be doing, is not about whether machines can think or not, but rather, if they can do what we as thinking entities do. And the answer to that question is clearly yes. The Turing Test is whether you can tell from a conversation with a machine if it is a human being or not, and if you can’t tell, then for all practical purposes it is human. You know what they say, “If it looks like a duck, walks like a duck, and quacks like a duck, then it is a duck.” Our latest computers can pass this test all right, no problem. In the same way we don’t need to bother with the nature of pain and other feelings. Like so-called “thought,” these are just side effects of brain activities and don’t actually do anything. As old man Skinner once said, we don’t run away because we are afraid. We’re afraid because we run away. Our whole focus is getting a machine to behave in the appropriate and rational way, not to feel anything, which is irrelevant, and we couldn’t do it anyway.”

  “I think I get that,” said Harry slowly. “I was kind of confused trying to figure out how you could ever make a robot think like a person, but I see where you’re coming from now. It doesn’t need to think, it only needs to act as if it was thinking.”

  “Precisely! We call it AI for effect, as opposed to AI for process.”

  “But won’t it create problems if it can’t understand how we think, or at least what kind of mood we’re in?”

  “Ah, but it will be able to effectively do so because it will have what we call a simulated emotional analysis engine. It will possess a whole array of sensors giving it the ability to read people’s facial expressions, body language, tones of voice, and so on, and it will be able to map the data it gathers to preconstructed tables that precisely describe the way a person who exhibits that behavior is feeling. Just because the robot can’t feel itself, it can still be a very good judge of how we feel, and in doing so it can read, and even predict, human behaviour. In fact, it will very likely do a better job than the average human because it can’t get confused by its own emotions, as we often can, and produce false readings.”

  “I think I could use one for my wife,” Harry mused.

  “The next essential point,” Vishnu continued as if Harry had not spoken, “is that the computers you’re familiar with are fundamentally deep down stupid. They are total idiots who would walk off the top of a tall building if you let them because they can’t learn and can only follow instructions blindly. But modern AI devices can use machine learning, of various kinds, so that they can be taught some basic rules by a supervisor, then go off and learn by themselves through the application of those rules. You may remember that the Deep Blue chess computer beat world champion Gary Kasparov in 1997. It did this by brute force, by calculating every possible move…”

  “But the word in the chess world is that IBM cheated to boost their share price.”

  “That could be, but more recently, the Alpha Go computer really did beat the human world champion of Go, a vastly more complex game than chess where brute force computing won’t work anyhow—there are too many possible moves. Alpha Go had to learn how to play all by itself, and it did, partly by analysing thousands of real-life games, and partly by playing millions of games against itself. So I can assure you that with that level of power and sophistication at our disposal, we should have no problem handling any of the relatively simple problems in business, law, economics, human relations, and so on that your robot will have to deal with. Let’s face it, people handle them every day and most of them aren’t very smart.”

  “So it looks as if my project is feasible
, then?”

  “Absolutely, no question,,” replied Vishnu. Harry was deeply impressed with the young man’s confidence, but then a potential problem occurred to him.

  “But if your AI computers are so much smarter than us and can learn just about anything, how are we going to stop them doing what they want? What if they decide to turn against us?”

  “Like Skynet and the Terminators?” Vishnu looked genuinely amused. “No, Harry, that’s a problem that first came up in the early years of robotics. The pioneers drew up some basic laws, and even though they were thought out a long time ago, they’re still valid. The First Law of Robotics is that a robot may not harm a human being or, through inaction, allow a human being to come to harm. Second, a robot must obey the orders given it by human beings except where such orders would conflict with the First Law. And third, a robot must protect its own existence as long as such protection does not conflict with the First or Second Law. Now these laws are embedded in the basic operating systems of all robots and they can’t be erased or altered by the machine. So you needn’t worry about robot rebels! And with the level of miniaturization we’ve now achieved and the huge strides in memory capacity we can jolly well fill your Genie of the Lamp with knowledge fit to bursting, with no worries!”

  Harry didn’t know what a Genie of the Lamp might be but he guessed it must be something pretty smart.

  “How about languages,” he asked. “Will the robot be able to handle them?”

  “Language competence and translation, in most cases, have reached a very high level, and your robot will be able to handle half a dozen languages or so with no problem. You might want to avoid Mandarin though. We’ve had some embarrassing translation slip-ups at international meetings, and I remember a business friend of mine who fancied himself as a linguist. Said to a Chinese opposite number what he thought was ‘I would like to set up a merger with your company’, but it came out as ‘I would like to have sexual congress with a duck’. The Chinese can be rather sensitive about that kind of thing.”