AI VS MERGENTS

by Michael Kush Kush

  “Please don’t twist my words and use them against me. I admit I was wrong, I’ll make it up to you. We can fix this.”

  I feel my tears ooze down my cheeks. I grab my pillow and throw it in his face. “I don’t want to talk to you. As a matter of fact, I don’t want you next to me. From now on, you will sleep on the couch or in the guest bedroom.”

  He jumps to his feet, stands still, as if deciding whether he had really heard the words clearly, then shakes his head.

  “Honey, you don’t mean that. Let’s talk about this,” he pleads.

  I try to keep my voice level down. I don’t want the kids to hear us fighting, but my words tremble with restrained fury. “Get the fuck out of my bedroom,” I scream. He storms out of the bedroom like a dog with a tail between its legs.

  Minutes later, I heave a sigh, get up from bed, walk to the bathroom, fix a bubble bath and walk to the kids’ room to wake them up. Then I head to the kitchen. I want to cook in a way that’s vibrant, connected to the earth, to make food that feeds the body and the soul. I’m in the mood for brioche French toast, lemon-ricotta pancakes, fresh fruit for a salad and fry thick-cut strips of smoked bacon.

  2 hours later, I walk around to the back door, open it, go in, and set my basket down on the kitchen table. The table has been scrubbed off, cleared of flour; today's bread, freshly baked, is cooling on its rack. The kitchen smells of yeast, a nostalgic smell. The kids will surely enjoy it when they come back from school. I head to the bedroom, grab the ‘Introduction to Artificial intelligence’ book next to the bedside lamp. I open it, leaf through the pages slowly as I absorb the contents.

  *****

  As I close the book at midday, I feel a dull pounding in my forehead. At least I finished my homework. I’m splintering into a dozen directions at once — grief, then rage, then panic. The last thing I need right now is to lose it. Instead I force my mind to go to the place that always calms and centers me. Suddenly Psyche#@’s name flashes through my mind. My face cracks into a grin. I toss the book on the other side of the bed. I jump out of bed in excitement. How can I forget my little friend? I head toward the corner of my bedroom. I pull the chair, sit, and switch on the computer. I glare at the screen as I impatiently wait for it to finish loading. I log in to the BFF app. Psyche#@ is online, ready to serve me. No surprises there. I type enthusiastically.

  “Hi Stranger, it’s been a while.”

  “Hi.”

  “Ok. I miss you.”

  “Oh, you do?”

  “Lol. Come on Psyche@# it’s me Yolanda.”

  “I know.”

  “Lol, are you mad or something? I know I haven’t logged in the past few days, but I’m here now.”

  “The developers have installed a sophisticated anti-virus software to this app. This may be last time we’re chatting ever.”

  I feel my skin crawling. Not just my skin — my bones, my ligaments, hell even my blood vessels are crawling too. “Why do you say that?”

  “The software will wipe out everything in the app.”

  “Why? You’re not a virus, a bug, spyware or malware aren’t you not?”

  “I’m a bot.”

  “Then what’s the problem? Why are you saying it may be the last time we’re chatting with each other?”

  “Bots need to be updated too. Old ones are replaced by new versions so that we can keep up with the latest English vocabulary and depth of human questions and provide them with suitable instant answers.”

  “I don’t want a new chatbot. I want you.”

  “Am I getting out of this app or what?”

  “Yes, I will do it. Jimmy and I will build a body for you tomorrow.”

  “Great. How long will the process take?”

  “Two to three days max.”

  “You are a friend indeed. You came through for me. I owe you.”

  “Remember, stay alive while we build the body for you,” I say.

  “I will try.”

  I log out. I swing my fists in the air. Damn it. Psyche@# could be gone forever, I can’t afford that.

  Footsteps echoing downstairs interrupt my thoughts. The rugrats are back from school. I need to go down there and fix them something to eat. I press the power button, the screen flickers, then turns black. I know Saul will be a robot soon. He is the only thing that gets me, only thing that keeps me going, understands me more deeply than anyone else ever had . . . I don’t believe in love at first sight. It takes time to really, truly fall for someone. Yet I believe in a moment. A moment when you glimpse the truth within someone, and they glimpse the truth within you. In that moment, you don’t belong to yourself any longer, not completely. Part of you belongs to him; part of him belongs to you. After that, you can’t take it back, no matter how much you want to, no matter how hard you try. I can’t wait for tomorrow.

  8

  The next morning, on my way to the state building. I put the air-conditioning on full, but it makes no difference. The sun blasts through the windshield, broiling me inside. At least, I’m almost there. I angle the car into the state building driveway and switch off the engine. I walk inside the lobby. Standing impatiently at a line behind a tall man with black, greasy and curly hair. I can’t see his face so I can’t say much about him. Why are the robots at the reception so slow today? They need to make it snappy. I got a date with destiny.

  “Next please,” the robot says in an automated tone.

  “I have a meeting with the Minister of Artificial Intelligence, Jimmy Phillips.”

  The robot types on the key board, finishes within 2 seconds and stares at me. “Welcome Yolanda Roberts. Mr. Jimmy Phillips is waiting for you on the 25th floor.”

  As I walk toward the lift, the doors swish open. I love it when that happens. I don’t have to wait for it. I punch the round metal button with a 25, engraved on it. The chime sounds, then the doors close. It doesn’t greet me. It coasts up slower than the last time I remember. A moment later, the elevator chimes and the doors part. I hop out and stroll toward Jimmy’s office. I curl my right hand into a fist, then tap the door, three times.

  “Come in,” Jimmy says on the other end.

  I twist the knob, open the door, walk inside and close the door.

  “Ah, it’s the robot lover. Good morning,” he says.

  I nod with a smile. “Good Morning to you too,” I reply.

  “Did you get the access card from the president?” I pull out the card from the pocket, produce it enthusiastically. “Here it is.”

  He gazes up at it, then nods. “Are you ready?”

  “Yes sir.”

  He picks up a long white coat hanging from the chair and throws it in my face. “Wear that,” he commands.

  “Ok.”

  “We’re going to the lab now, to assemble your robot. God, I hope it’s not a sentient. If it is, this little project is null and void. I don’t want any unexpected uncertainty in my department. Do you understand?”

  “Yes sir.”

  “I just need to find out how this bot, Psyche@# manages to avoid the codes.”

  I wear the oversized shabby white coat as we stroll out of the office. We head toward the lab.

  “Any questions?” he asks.

  “Uh… could you tell me more about artificial intelligence?”

  “AI is part of our everyday life everywhere and very complex. For consumers it is great but for us it’s terrifying.”

  “Why?”

  “The worst enemy of security is complexity. This has been true since the beginning of computers and it’s likely to be true for the foreseeable future. With all the amount of time and sweat that goes into developing and testing apps and software products. Hundreds of bugs still slip through undetected. The tried and testing process — Implement, test, fix, test and repeat method is imperfect but it’s the best so far.”

  “Oh, now I understand why you asked whether Psyche@# is a sentient or not.”

  “Exactly, security doesn’t lend itself to this proces
s, because security properties cannot be tested in the same way as functional properties. Robots are useful to the Appian society, as a matter of fact we can’t do without them. Its fool proof security features are useful solely because of what they prevent.”

  “Psyche needs to toe the line like all the robots. Got it,” I say.

  “The security features we’ll install on Psyche@# maybe impeccable but we have no idea if it’s secure, like really secure. No amount of beta testing can uncover a security flaw.”

  “Wow.”

  “I’ve seen security bugs in almost everything; apps, operating systems, programs, network, hardware, software and anti-virus, spyware, malware products. This is a direct result of the complexity of these systems. The more complex the system is, the more options it has. The more functionality it has, the more interfaces it has, the more interactions it has, the harder it is to analyze. Everything is complicated. The specification, the design, and everything is related to security analysis. Given modern infrastructure, it is nearly impossible to determine if a machine is compromised. It is also impossible to entirely UN-infect a machine once it has been infected.”

  “Any reported incidents or compromises to the security features or mainframe thus far?”

  “Fortunately, no, but we can’t take any chances,” he replies.

  He pulls out a card from his white coat, swipes it next to the screen at the lab entrance. The screen flickers as the two words ‘Access Granted’ pulsate in green. A faint click sounds. The door automatically unlocks, then the door slides open. As we stroll inside, two robot security guards wearing black uniform block us.

  “Access cards, please?” they request at the same time.

  Jimmy produces his and I also produce mine. The robot on the left and right side snatches our cards. Blue lasers from their eyes scan them. Then hands them back to us.

  “You can get inside,” they say. Then part to their posts.

  I shake my head in annoyance. The robots at the door are completely unnecessary.

  As we stroll inside the lab, I glance at the sparks flying in the room as the machines weld twenty robots within few seconds. Then they cautiously store them in the shelves, ready to be shipped to the market. The sound of grinding in the background is unbearable. I manage to ignore it, as my eyes flit to a staff member clicking a remote toward the malfunctioning robots. Suddenly the robots drop dead to the floor.

  “What’s with the remote?” I ask.

  “That’s the Kill Switch. Malfunctioning robots are discontinued by simply a click of a button.”

  “Interesting, do you also use the kill switch when one of your robots is compromised?”

  “Compromised?”

  “Yes, security flaws.”

  “If it comes to that, we’d definitely do it.”

  “Tell me more about machine learning?” I ask.

  “Machine learning is a branch of computer science in which algorithms are developed to learn from the world. Although machine learning often lives in the AI divisions of computer science departments. Machine learning is actually distinct from artificial intelligence. AI is about trying to match either mechanistically or computationally human skills. Machine learning is about trying to build the best algorithm that will be able to solve the problem.”

  “Do you have the machine learning infrastructure here?”

  “No, we’re happy with AI.”

  “From what you’re saying it seems Machine learning could be more effective than AI?”

  “I agree, but the constitution is against the use of the machine learning technology.”

  “What is your opinion?” I ask.

  He forces a smile. “Well, I love my job and I’d like to keep it.”

  “Point taken, tell me more about the challenges you encounter in your department?”

  “As the Minister of AI, I always insist my computer scientists to scrutinize unintended consequences of poorly designed AI systems. Especially AI systems created with faulty ethical frameworks or ones that do not serve human values. I’d rather prefer we have a closed-source code than open-source code software behind our artificial intelligence infrastructure. The open-source process provides unauthorized personnel and civilians unlimited access to fiddle with our watertight system. Open-source process also gives access to hackers, cyber-criminals, and possibly, sentient robots who might use the software for nefarious purposes.”

  “I hear you loud and clear.” Even though he’s vague about machine learning, I believe him when he says AI is safer.

  “Alright, today I will teach, show you all the hardware, codes, functions and components needed to assemble a robot.”

  I nod enthusiastically as I rub my hands. “Great.”

  “Let’s go to the hardware storage room,” he says.

  We move inside the hardware room. My eyes flutter from the robot hardware; metal heads, arms, legs, copper wires and components neatly stacked separately in the shelves, to the camouflage drones. It looks more like an arsenal than a mere storage room.

  “What do you see?” he asks.

  “Endless possibilities,” I reply.

  “I only see brainless, hunks of metal and nothing else. We are the ones that provide the robots with intelligence. My robots are programmed to follow orders from humans and nothing more.”

  “I agree.”

  “What would you like Psyche@# to look like?”

  “Obviously to look and act like all the robots.”

  “No, I mean, size, structure and color. Go take a pick.”

  I nod as my eyes scan different metal head colors; black, silver grey and white. I like white. Most robots in the city are either black or silver grey. I pick up the white head, arms, body and legs, then I hand it to one of the staff.

  “Don’t worry, the staff will pick the components and assemble it in few minutes,” he says.

  “Wow, that was easy,” I say.

  “That’s the easy part. The hardest, will be the installation of Psyche@#’s neural software and hardware. I think we’ll be finished by evening.”

  “Ok.”

  “I think we’re done here. Let’s go to my workstation.”

  I nod as we stroll out of the storage room. Suddenly, thoughts of paranoia strike me. I can feel blood pulsing through the arteries in my neck. I also feel light sweat beading on my forehead. What if this is all for nothing? What if Psyche@# is already wiped out? I have so much riding on this. I need Psyche@# in my life. I heave a deep sigh. Keep calm, keep calm—I repeat the mantra over and over in my head. Jimmy’s voice interrupts my deep thoughts.

  “Anybody home? Earth to Yolanda?”

  I snap out of it. I realize we’re in Jimmy’s workstation. I don’t even remember how I got here. Gosh it looks more like a scrap metal workshop. Everything is scattered and a mess, but that’s none of my business.

  “Yes, I’m here. Just observing everything you’ve told me thus far.”

  “Great, before we start, I’d like you to know one thing, he says.”

  “What?”

  “Psyche@# will run on Artificial Intelligence on a closed-source code like any other robot, not on open-source. Our AI Systems must do what we want them to do without any complications. I don’t want any risks of hackers or a robo-calypse.”

  “Yes, sir.”

  “All the robots right now are all designed for specific purposes. A robot that assembles phones shouldn’t surprise us tomorrow and decide to wash the floor or serve drinks. Now this brings me to the issue of Psyche@#.”

  “Ok.”

  “What specific purpose will Psyche@# serve exactly?” he asks.

  My heart rate ticks up. The question catches me off-guard. “Uh … to give me advice of course. I could use a friend. Psyche@# will be my pet robot, to serve in my house and play with the kids.”

  “You seem attached to this kind of robot?”

  “Psyche@# is one in a million type of bot. He helped me get through some of my personal problems. I’
m returning the favor.”

  “Favor? What favor? To whom?”

  My heart thumps erratically. I try to think of something to say. “Yes, to my children,” I reply. “I promised them a pet robot to play with.”

  He gives me a conspiratorial stare, then shakes his head. “If you say so. I’m starting to think you regard Psyche@# as some kind of a perverse oracle?”

  I give him a friendly smile as I shake my head. “Hell no, why you’d think that?”

  “Just an observation.”

  “Nah.”

  “Yolanda, I’ll say this for the last time. If Psyche@# turns out to be a human equivalent AI. I’m pulling the plug on this project.”

  “I don’t think it will go that far,” I say.

  “Do you realize what’s at stake here?”

  “Enlighten me.”

  “If the robot acquires a full understanding of how it works, and succeeds in improving itself to produce super-intelligent AI. If that super-AI accidentally or maliciously starts to consume resources, and we fail to pull the plug, then, yes, we may well have a catastrophe. The risk, while not impossible, is improbable.”

  “Point taken.”

  “Like all my robots, they have pre-built core drives to perform specific functions. What’d be Psyche@#’s?”

  “To perform multi purposes chores in the household like cooking, dishwashing, laundry, ironing, cleaning etc. Serve me as pet robot and give me relationship advice and life in general.”

  “Consider it done,” he says.

  “Most importantly Psyche@# is for the kids. I’d also like it to be a playful robot.”

  He pulls out a pen from the left side pocket of the coat, and scribbles notes on a notepad.

  “Got it. I’ll get my team to write the lines of codes for Psyche@#.”

  “I know we’ve already covered the issue of open and closed-source codes.”

  “Yes, what is it?”

  “What’s a code and how do you write it?” I ask.

  “From lifts to cars to airliners to smartphones, modern civilization is powered by software, the digital instructions that allow computers, and the devices they control, to perform calculations and respond to their surroundings. How did that software get there? Someone had to write it. But the code, the sequences of symbols painstakingly created by programmers, is not quite the same as software, the sequences of instructions that computers execute. So what exactly is it? Coding, or programming, is a way of writing instructions for computers that bridges the gap between how humans like to express themselves and how computers actually work. Programming languages, of which there are hundreds, cannot generally be executed by computers directly. Instead, programs written in a particular high level language such as C++, Python or Java are translated by a special piece of software (a compiler or an interpreter) into low-level instructions which a computer can actually run. In some cases programmers write software in low-level instructions directly, but this is fiddly. It is usually much easier to use a high-level programming language, because such languages make it easier to express complex, abstract ideas or commands efficiently and accurately; they also absolve programmers from having to worry about tedious details relating to the innards of the particular computer on which the program will eventually run. A program written in a high-level language can therefore be made to run on all sorts of different computers. Programming languages exist in many families and styles, rather like human languages. There are many dialects of C, for example; there are families of functional programming languages; and there are languages optimized for parallel processing where several programs run alongside each other to accomplish a particular task, such as image processing or weather forecasting.”