Uncanny Valley

I had to be in the lab at eight in the morning, which, after closing night of “Gunder’s Hollow” and cast party and getting awoken at four by my mother’s hologram, felt a little like death. But, I had my coping strategies for just such mornings as this one: namely, Lavazza’s triple shot latte. The little coffee shop was around the corner from the psych lab where I conducted my experiments, but test subjects would arrive at any minute. I prayed for no line.

Thank you thank you, I breathed—apparently I’d arrived right during the morning lull. The bot behind the counter asked my order and busied itself with her four arms, ringing me up and making my drink at the same time. She barely had to move.

Within two minutes, I held my brain function in my hands, slurping so it didn’t scald my throat on the way down while half-walking, half-running toward the lab.

“Hi guys!” I greeted the line of human volunteers queued outside the door. The medic bot waited patiently as well, wearing a cute little white hat with a red cross on it to identify her. “Thanks for your patience.” I doffed my backpack and jacket, pulling my handheld interface out of my backpack to collect signatures from the participants. They had already sent me their consent forms—I only needed to match fingerprints to participants, which we used in lieu of A.E. chip tracking for identity verification because it kept costs down. I gave the interface to the first girl in line. While they queued up to scan their fingerprints, I hurriedly unlocked the door to the waiting room, turned on the lights in there and in all of the actual testing rooms, and bid the medic bot to follow me.

“We can set up for the blood draws in here,” I told her, pointing to a little counter at the front of the waiting room, designed for a receptionist. I should have been sitting there myself, checking everybody in… and I would have been, had I woken up about an hour earlier.

“Very good, Miss Cordeaux,” said the bot in a tinny voice, rolling over and distributing her instruments. She was a funny little thing: about half my height, but not as short as my own bot Madeline. She had no dexterity at all in her lower half, since she had no legs. But her arms were extendable just in case she needed to pick something up off the floor, and her hands were as supple as a surgeon’s. She had clearly been built for her purpose.

I left the medic bot to set up, and went to make sure the rooms were ready, gripping my latte for dear life. Pulling up my notes, because I could never quite remember all the steps, I went through them one at a time: interface on, Artificial Experience images and videos specific to each volunteer ready to go, matched to the rooms where each volunteer will be… Then there were a series of commands I didn’t understand but copied very exactly, in order to allow the individual’s A.E. chip to image the participant’s brain and send the images to the VMI (Virtual Magnetic Imaging) for analysis, alongside a snapshot of the image that participant was seeing at the time.

When I’d finished with the last room, I breathed a sigh of relief, and parked my butt in the receptionist’s chair in the waiting room. Beside me, the medic bot had just finished taking blood from the first participant, the young girl to whom I’d given my interface at the front of the line. She was about my own age; clearly she only volunteered for the money. Not that I blamed her.

I consulted my list. “Carrie? You’re in room one. The A.E. is all queued up for you. When you go in and sit down, the lights will dim and the experiment will begin.”

Carrie nodded at me, looking like she could use a latte too.

The medic bot and I repeated the process for the next seven volunteers, until the room emptied out. Once the medic bot had collected and labeled baseline blood samples for each participant, she announced to me, “I will analyze these for your neuropeptide salheptonin, and report back to you.”

“Thank you,” I said, and she wheeled away.

I was alone. With a heavy sigh, I plopped my forearms on the counter and my head on my forearms. To my surprise, I didn’t feel all that tired anymore though—the espresso had done its work.

I have an hour, I thought. That’s how long the first phase of the experiment took. I could of course log in and see the VMI brain images and the artificial experiences that had stimulated them in real time, but frankly… I wanted to work on my novel. I had an hour, after all.

I pulled my notebook out of my backpack, along with a pen. I could write the whole thing on an interface or even dictate it, of course, but I liked old fashioned things wherever I could get them. Hence the analog clock in my dorm room. I liked to visit archaeological sites, especially old castle ruins, where I’d try to imagine what life might have been like in the Second Age, before Synthetic Reasoning and Artificial Experience and even the labyrinth. Maybe even before light bulbs.

I skimmed the pages of what I’d written so far, trying to get back into the flow of it. The story followed Elizabeth, who grew up in an orphanage controlled by the king. The king’s son, Nikolai—devastatingly handsome, of course—grew up visiting the orphanage because his father wanted him to maintain a connection to the people, so he develops a sort of friendship with Elizabeth. She grows up and becomes the handmaid to the princess in a neighboring nation, and Prince Nikolai still comes to visit often. I threw in some intrigue about the mob and Elizabeth being whipped for her impertinent familiarity with royalty. But Nikolai is perfect and wonderful and doesn’t care about her rank. Elizabeth can’t believe he doesn’t care, so she shuns him. Eventually he starts to believe she really doesn’t have feelings for him, and he actually considers the Princess of Spain whom his father wants him to marry. Elizabeth looks on, devastated…

“Whatcha doin’ there?”

I jumped, and gasped, a hand flying to my chest as the other covered my pages as best I could. “Geez! You scared me.”

“I see that. You might wanna lay off the espresso, I think it’s not good for your nerves.” Liam, the post-doc I worked under, raised an eyebrow at me, his lips curled in that characteristic smirk of his. I shut the notebook immediately. “What are you writing?”

I shrugged. “Just killing time,” I told him evasively. If he knew, I’d never hear the end of it.

“I guess it’s too much to hope that you might be brainstorming new experiments…”

“No, just…” I decided it would be best to change the subject. “What are you doing here, Liam?” He usually didn’t come by during my experiments—human psychology wasn’t his thing. He waited for me to bring him conclusions, and then translated them into algorithms. I’m technical—I don’t do people, was the way he usually put it. It was sort of tongue-in-cheek though, since his social skills were fine, as far as I could tell. Although he was undeniably eccentric.

He took a deep breath, his face suddenly serious. I knew what he was going to say before he said it—only one subject rendered him so somber. “You heard Halpert’s address?”

I nodded. I had watched it after Mom called last night, even though I’d just sent her a quick comm about it and gone right back to bed after that.

“This is bad,” he said, in a tone of voice that prompted agreement.

I opened my mouth and closed it again, waiting for him to elaborate. Which he would, if I didn’t do it for him. He couldn’t help himself. “You’re afraid that if the bots get creativity…”

“They’ll become superintelligent, way surpassing humans!” he finished. “Once they become creative, they can devise ways to make themselves smarter, without our intervention at all… and that’s an exponential curve. This was exactly what the Council of Synthetic Reason was supposed to prevent! I wrote a manifesto about it on my locus as soon as he’d finished and issued a counter-call to Halpert’s. We have to band together across the globe and hound our senators to stop the challenge, or else this collaboration will lead to the extinction of the human race—”

“Oh, Liam,” I rolled my eyes. Among his many eccentricities, Liam had a conspiracy theory labyrinth locus with apparently millions of followers, or so he was always telling me. He was a sort of B-list celebrity in that way—in pseudonym only, since I never saw his picture on there anywhere, and he didn’t use his real name. “I don’t want to become a target!” was his explanation for this. I’d gone to his locus once before, expecting to see skulls and crossbones and apocalyptic imagery, but I had to admit I was grudgingly impressed. It looked classy enough, and the one post I’d read was remarkably well-written, cited, and devoid of groundless assertions. I also saw that it had thousands of shares, and had only been posted that morning.

“It’s true, if bots gain the creative ability to problem-solve like we have, one of the most probable futures is our extinction!” Liam insisted. “Don’t roll your eyes at me!”

“You’re such an extremist. It’s quite a leap from giving bots creativity to the extinction of the human race.” I emphasized each word, trying to make him recognize how silly he sounded.

Undaunted, Liam raised his index finger as he counted off his points. “One: bots are single-minded. They exist only to fulfill their programmed core purpose. Two: they are ruthless logicians. Within the parameters of their abilities, they perform that core purpose as efficiently as possible. Three: at present, they can only learn to do procedural tasks, they cannot yet solve complex and interconnected problems. Once they have creativity, though, they’ll be able to do that just like we can. So what happens if you have a bot who has creativity, and whose goal is to make itself smarter? It does nothing else, 24 hours a day, seven days a week. Trial and error, trial and error.” He demonstrates a logarithmic curve with his hand, eyes wide as he shook his head at me for emphasis. “Exponential growth!”

I stared at him, unimpressed. “Okay fine, so they become superintelligent. How does that lead to our—”

“Because!” he cut me off. “Maybe the bot who first becomes superintelligent has no interest in us at all, but it is interested in spreading its own upgrades to other bots. Now every other bot in the world, with every other possible core purpose, has infinite creativity at its disposal in order to fulfill its purpose! Just take one of those bots: let’s say one of them manufactures shoes, okay? Its goal is to manufacture as many high quality shoes in as short a period of time as possible. But it’s ruthlessly logical, and narrowly focused. What if it decides that humans are using too many resources that it needs to make its shoes? Logical solution? Kill all the humans! Then it can mine all the resources on earth and the moon and Mars and any other planet they colonize after we’re gone, and voila! The entire universe becomes a giant shoe closet!”

I cracked up a little. I couldn’t help it. “All right, I see your point that they might carry out their core purposes in unexpected ways. But it seems like we should be able to safeguard against that…”

“You mean, predict all the possible ways that a superintelligence leagues beyond human understanding might interpret their core purposes?” Liam countered. “How would we do that, exactly? That would be like an ant trying to fathom the mind of a man! The only possible way to approach this is to stop it before it can happen. QED,” he added with a triumphant flourish, sitting back in his chair. He always said ‘QED’ when he felt like he’d made an unassailable point. All I knew was that it was a math term in Latin, and it basically meant the same thing as, “infinity times, the end.”

I looked up before I could shoot back something sarcastic at Liam, sensing that we were not alone. Carrie, my first test subject, stood meekly in the door frame. I saw her mouth fall open a little when she saw Liam, and she colored a bit. I knew why: Liam was your classic tall, dark, and handsome, with blue eyes and that angular jaw that drove women crazy. It annoyed me. All right, he’s good looking, people. Get over it. If they knew him the way I did, it wouldn’t take long.

“Done?” I said to Carrie, leaning past Liam with a firm smile and forcing her to look at me as I pointed down the hall. “The medic bot will take one more blood sample in the next room over. Then come back here and I’ll get you paid.”

She vanished, and I turned back to Liam. One thing I had to give Liam: he never seemed to get that self-complacent air that most attractive men get when they find themselves admired. I’m not sure he even noticed.

“But there might be one thing we can do as a potential safeguard, just in case Halpert’s challenge is answered…” Liam went on, as if the interruption hadn’t happened. “Program them with a moral code.”

“You’re talking about Isaac Asimov’s Three Laws of Robotics?” I raised my eyebrows, trying unsuccessfully not to smirk. I only knew these existed in the first place because of something I read on Liam’s locus, and I couldn’t exactly recall what they were—something about the hierarchy of robotic priorities as a way to keep them from becoming dangerous to humans.

“That’s fiction,” Liam scoffed. “This is reality. ”

I shrugged. “Okay. So program them with a moral code, then. What do you need me for?”

“I need you,” he leaned forward, his eyes staring melodramatically into mine, “because everybody thinks we won’t get creativity except through emotion. They go together, peas in a pod.” He crossed two fingers together to illustrate. “Humans have a moral code too, but when we break it, why do we break it?”

“Emotion,” I sat back in my chair, catching his drift now.

He nodded. “So let’s recap.” He held up his fingers one at a time. “One: Superintelligent bots with unlimited creativity. Two: Emotional superintelligent bots… who can presumably therefore overwrite or ignore their programming if they don’t happen to like it. So sure, we could give them a moral code, but who’s to say they’ll follow it?”

“If they can overwrite their programming, maybe they decide they don’t want to manufacture shoes, then, either,” I pointed out. “Maybe they want to… I don’t know, optimize the fertility of giraffes instead.”

Liam shrugged. “Maybe, but both will be equally fatal to us—too many humans means not enough space for giraffes, either. Point being, we need to instill at least some semblance of morality in the bots that they won’t just overwrite at will. Maybe they can, but they’d choose not to. So what does that take? Empathy or something? I don’t know, this is your department. I’m just speculating.”

I bit my lip, eyes absently scanning the air as I tried to mentally retrieve what I’d studied about this. “Empathy comes from mirror neurons, which cause us to share the emotional experience of another person. Emotion comes from the limbic system, and it’s mediated by a whole lot of neurotransmitters…”

“But sociopaths don’t have emotion or morality, right?” Liam cut me off. “Seems like they’d be a good group to study, to see what’s missing…”

“They have emotion,” I corrected, “they just lack empathy.”

“Perfect!” Liam said, “that’s exactly what we’re looking to avoid: a race of superintelligent sociopaths! So you’re on that. Design me a human experiment to define the relationship between emotion, empathy, and morality.”

I laughed, incredulous. “Oh, is that all! What about my thesis?” I gestured at the lab where we currently sat.

“You’re already studying human emotion, aren’t you? Something about that new neuropeptide?”

“Yes!” Then I reminded him, since I knew he wouldn’t recall the particulars, nor would he care, “We’re shooting VMI images of the brain when participants see an image of someone they love, and then drawing pre- and post-blood samples to see if salheptonin increases simultaneously. My theory is, it’s the neuropeptide responsible for the physical sensations of desire.”

“Okay, so this is just a pivot,” Liam shrugged.

“This is a completely new project!” I protested. “And it’s enormous, by the way! If I ‘pivot’ like you’re suggesting, someone else will complete my research before I do, and I’ll have to start all over with something else if I want to graduate!”

“Rebecca!” He made exploding movements with his hands. “End of the world! Who cares if you have a degree if the entire human race is extinct?”

I shook my head, incredulous. “I hate you.”

“No you don’t,” he flashed me a rakish smile. “Besides, like you’re having so much fun working on this thesis of yours, anyway?” he gestured at the stark white room in which we sat, and at the notebook I’d closed. “Every time I come upon you unannounced, you’re writing the next Gone with the Wind. Or reading a novel from the Second Age. Or humming to yourself to learn the harmony for your next musical performance—you never told me when the performance was, by the way.”

I flushed. “Just because I have lots of interests—”

“Hey, hey. It’s okay,” Liam cut me off, hands in the air with an expression of mock seriousness. “I’ve got millions of people following my labyrinth locus who agree with me. I don’t need you to be one of them, although it would be very nice. But you remember what Dr. Yin said about you in your review?”

I sighed, and quoted Dr. Yin: “‘Technically very competent, but her heart does not seem to be in her work.’” I hadn’t known whether to be flattered or insulted when I read that.

He nodded. “Yeah. And you know what she said to me privately? She asked if you were planning to go for your Ph.D., because if you were, she’d like to offer you a spot in her lab. You’re that good, Bec. And considering you’d rather be doing something else, in a way that’s even more impressive. Who knows, this project might even turn into your Ph.D. thesis!”

“Eh-hem.”

I looked up, startled. One of my middle-aged male test subjects stood in the door frame, ready for his exit blood draw. Carrie queued up behind him, waiting to get paid. I sent the man to the medic bot and beckoned Carrie forward, scanning her fingerprint again to deposit her compensation and trying to ignore the way her eyelashes shyly fluttered at Liam.

Liam never complimented me. He usually restricted his communications to fluent sarcasm punctuated by obsessive rants. It’s only because he’s buttering me up to do something for him, I thought as I dismissed Carrie.

When I’d finished, Liam resumed his unnerving stare deep into my eyes, triggering involuntary heat in my cheeks. I wished he’d blink, or look away occasionally, or something. He went on, “What I’m saying is, I need someone on the human psych end to help me program a morality the bots will actually follow, as a safeguard in case Halpert’s challenge does succeed. At least then, even if superintelligence does emerge, we’ll have ‘warm and fuzzy’ superintelligent bots, with a soft spot for humans!” He mimed his impression of ‘warm and fuzzy’ as he said this, which turned out to look a lot like jazz hands. I shook my head at him with a suppressed smile, and his eyes twinkled at me in response. “Once we have your succinct definition, then Nilesh and Larissa and I can reverse engineer it in mathematical terms and make it into an algorithm. Then you can go back to working on whatever you want, I promise.”

“And if someone answers Halpert’s challenge before we get the chance?” I asked. “Because what you’re asking me to do is borderline impossible, I’m just saying…”

Liam shrugged, unconcerned. “Then we’ll kill him, and bury his research before anybody else can use it to buy us more time. Please be quick though, would you? I’d like to kill as few people as possible.”

“Glad to hear you have a backup plan,” I nodded, trying to match his deadpan delivery without success.