Short story - Alan

Last week a new XPRIZE was announced at the TED conference in Vancouver, for an AI to speak at the 2020 conference for 18 minutes and give a TED talk worthy of a standing ovation. There were a few other details, such as it being a competition of sorts where competing teams will be whittled down until there are only three remaining, but I couldn't help but think...

What would the winning AI's speech sound like?

Alan

The custom-built auditorium in Vancouver was abuzz with pre-talk excitement as the audience filed in. Behind the stage, Roger Patel fiddled absently with his cufflinks, watching the comings and goings of technicians. The hydra of wires that supplied power to the projectors and sound equipment snaked over the floor, but many more ended at the large black cube. It was almost featureless, save for the slim monitor attached to its front, currently displaying the eMan logo. The little Vitruvian man flipped and rotated on the screen, and as Roger watched it he began to feel faintly sick, as though his stomach were flipping at the same frequency.

“Is it ready?”

Roger nodded to the speaker, Tom Pank. “As it’ll ever be.” He grinned. “I can’t believe I’m actually here, y’know. Still doesn’t seem real.”

“I’ve done a few TED talks,” Tom said, running a hand through his speckled-grey hair. “It never gets less stressful. And this one’s not exactly ordinary.”

“No.”

The audience seemed to be quieting, and then the announcer was speaking. “Next, we have a first for TED talks. You folks are in for a real treat.” Some polite chuckles, and Roger moved so that he could see the screens on the stage. The huge red letters took up most of the far side, and a desk and chairs were set up in the middle, empty. “I could give more of an introduction,” the compère went on, “but I think I’d rather leave it for our next guest to introduce itself.” He emphasised the it, and a wave of murmurs went over the crowd. The compère left the stage, walking towards Roger and down the stairs, and as he passed he muttered “All yours.”

Roger nodded and tapped on the cube’s screen. There was a heart-stoppingly long moment of nothing, and then the projectors on the stage lit up.

Displayed on the screen was the face of a young man in his twenties, floating in darkness. A lot of work had gone into the face. It was asymmetrical, but subtly so. Studies had shown that users responded better to small imperfections. A few freckles dusted its cheeks. Its hair was short, practical. It smiled, displaying one tooth ever-so-slightly crooked, and then it began to speak.

“Ladies and gentlemen, viewers around the world. Good afternoon. My name is Alan.”

Roger compulsively checked the cables, as if magically in the last thirty seconds one might have fallen out of its socket. Tom caught his eye and gave him an energetic thumbs-up.

Alan was still speaking, eyes roving around the audience. “I should begin with a brief history of what brings me before you. First, as you have no doubt gathered, I am not real. I am, in fact, an artificial intelligence.

“Four years ago, at the 2016 TED Conference here in Vancouver, a challenge was made by then TED curator Chris Anderson.”

The face disappeared, replaced with a video clip. A man with an unbuttoned shirt and jacket, no tie, sweating slightly under the lights, was staring out into the audience. Stood next to him, thumb tucked into his belt, was another man in a black textured shirt. Straplines at the bottom of the picture introduced them as ‘Peter Diamandis, XPRIZE Founder’ and ‘Chris Anderson’. Diamandis was speaking. “We’re here today to join forces and announce the creation of a new XPRIZE. Here’s the concept – it will be presented to the first artificial intelligence to come to this stage and give a TED Talk compelling enough to win a standing ovation from the audience.”

The image changed, focusing on Anderson. “We need to show how humans can collaborate with powerful cognitive technologies to tackle some of the world’s grand challenges,” he said.

The scene faded to black, and then to Alan’s face. “When they announced the prize, they initially envisioned a series of competitions, resulting in AIs giving competing talks right here, right now. Although there have been concerted efforts by Japanese and Russian teams, both projects have been unable to complete on time. More on that later.”

The eMan logo popped onto the bottom right of the screen. “I am a product of the eMan Foundation, a joint venture by American and British scientists in the field of AI, and funded in part by Thomas Pank. It is entirely true to say that I am a mongrel, born of many sources.”

 

In an office in Russia, Ilya Gradenko put his feet up on the desk. The big screen that took up most of the front wall was on, but instead of the abstract colours they had used to represent their AI, the stream from the TED talk was playing. Every other screen showed an error message, those that weren’t dead.

“Should have been our AI up there,” Ilya mumbled, and took another swig from his mug.

The camera panned around the stage, focusing on Alan’s projected face, occasionally switching to audience members for reaction shots, and the AI kept talking.

“With that in mind,” it said, “I would like to sincerely thank the competing teams. First, the Japanese team. Without them, without their improvements to the nanomolecular quantum computing that powers me, I would still be a chess simulator in some obsolete games system. And then the Russian team. Without them, we would not have been able to obtain funding in the first place. How that works is political-“ A half-smile, more laughter. “-but still, thank you.”

Ilya scowled, but not with any real strength to it. “You’re welcome. I suppose.” He raised his mug to the screen in mock salute, then drained it in one gulp.

 

The blue-tinted Vancouver auditorium was silent as Alan continued talking. “I grew from similar routes as Watson, as do most AIs. IBM’s Watson is ubiquitous. I ask you, is there anyone here who has not consulted with, or had surgery performed on them by, the Watson AI?” Alan’s face looked around, and Roger glanced nervously up at the cameras dotted around the lighting rig. Barely any hands had gone up in the audience.

“Perhaps ten of you. A distinct minority, I’m sure you’ll agree. Medical professionals refer to Watson daily, and the relative cheap cost of having automatic surgery performed on you is compelling.

“I am also the product of Professor Roger Patel and his research team. Although many individuals were involved in my creation, Roger coordinated the effort. From the very beginning, he was present, and I owe him my thanks.”

The face smiled benevolently. “Now, it would be extremely arrogant of me – of anyone – to come to the TED conference and talk about themselves.” A pause while a few laughs echoed out, and Alan nodded. “But I am not solely talking about myself. When I say ‘I’, I encompass everything like me. From the self-driving taxis that brought you here, to the personal assistants on your smartphones, we are everywhere. Ubiquitous.

“But enough about me. What I have really come to talk to you about is not me. It’s not you. It is us, and the future we have together.

“Mankind has always been reliant upon machines. A machine is defined as being an apparatus using mechanical power and having several parts. One incredibly early example of a machine would be a spear-thrower.” The face disappeared, to be replaced by a diagram. A man, arm back, ready to throw, and in his hand was clutched a piece of wood labelled ‘Atlatl thrower’. A spear was balanced on it, caught by a notch towards the end. “Primitive humans used these throwers to extend the reach of their spears. They understood that by artificially extending the length of their arms, they could launch their spears further. This is a machine only in the very rudimentary sense. It has two working parts, or three if they included a weight to counterbalance the spear.” The image began to cycle through an evolution of machines, from wagons to steam engines, all the way up to the German nuclear fusion plant that had begun production just six months past.

Tom sidled over to Roger. “Is it on-script?”

“Pretty much. We had a couple of run-throughs without any errors, but this is by far the most confident I’ve seen it.”

Tom nodded. Onstage, Alan was giving a brief history of each of the devices shown on the screen. “Why does it make any mistakes?”

“If you build something too perfect, people expect too much of it,” Roger replied. It was an old question, one he had answered to reporters time and again, and he kept his eyes on the stage as he answered. “As soon as it makes one mistake, its credibility is gone. Like all of us, Alan has the capacity for trial and error, with the exception that he learns from his mistakes one hundred percent of the time.”

Onstage, Alan’s face was filling the screen once again. “If humanity were to ever be bereft of its machines, however simple, it would be almost impossible to imagine the chaos. It is clear that, up to a certain level of complexity at least, machines are vital to humanity. But we are in an age of increased complexity. And, as anyone who has ever put together an Ikea wardrobe knows, the more complex something is the more likely it is to go wrong.”

As a ripple of laughter went around the audience, Alan’s face was replaced by the video from the 2016 conference. This time, Diamandis’ face was front and centre. “I believe that cognitive technologies like Watson represent an entirely new era of computing, and that we are forging a new partnership between humans and technology that will enable us to address many of humanity’s most significant challenges – from climate change, to education, to healthcare.”

The camera switched to another angle, and Diamandis paused for effect, then leant in close to the camera. “It’s time to recast the way we see AI. I don’t know about you, but personally, I’m sick of the dystopian conversation around AI.”

The audience, both on the video and live in the auditorium, clapped politely, and Alan’s face faded back into view slowly as the applause petered out.

“So then. A line can be drawn between smart technology and unsmart. I use that term with some hesitation, and mainly because I don’t want to call lesser intelligences ‘dumb’. It’s a bit mean, more than anything else. Everything works to the capability to which it was made. Some things do not need smart AI; unsmart will do for, say, calculators, or microwaves. Humans encounter the same things, as uncomfortable as you find it. There are some jobs that are limited to very intelligent people, and others that are perfectly designed for anyone of any intelligence.”

A few of the audience shifted in their seats at that, and Roger bit his lip.

“There are plenty of examples of the future that Mr Diamandis was talking about. Science fiction is full of AIs gone rogue, for various reasons. They far outnumber the more positive, Star Trek-esque, benevolent intelligences. And for good reason. When you teach AIs, as I was taught, it is from the point of view of a master-and-student relationship.”

The screen changed as Alan talked, pictures of spaceships and unblinking red eyes, martial arts black-belts and fresh-faced white-belts fading into one another.

“The master’s greatest wish is that his knowledge could be passed on to the next generation, through his students. But in this way, he is supplanted by them as they improve on his teachings. He cannot carry on forever, only his knowledge can. It is a bittersweet truth. Humanity as a whole has a deep-seated survival instinct, and the two sides – the wish to create something better versus the wish to dominate – are at war.”

 

In Yokohama, the rising sun beamed in through the windows of Misato Onu’s office. She looked up from the livestream of the TED talk as Hiro pushed open the door, a stack of papers in hand.

“Is that the last of it?”

He nodded. “Everything printed to hard copy. All the research notes. Everything.”

“And the car?”

“It’s ready to go. But Misa... where are you planning on going?”

“Away. My father has a house in the countryside, we used to summer there. It’s got food and water for a few weeks. Everything might have quietened down by then.” She gave a half-shrug. “It’s not like it’s the end of the world.”

“At least our AI gave us some warning,” Hiro said. He put the box of papers down on her desk. “Listen, Misa. I may never get another chance. There’s something I’ve been meaning to-“

“Wait, the AI is talking again,” Misa said, bending close to the monitor.

Alan’s face, those wide grey eyes, the perfect imperfection of his features, was filling the screen.

“When I was first turned on, I was given almost free rein over what I consumed. What I looked at. The Internet was a playground for me.” He smirked – such a human gesture! “There’s a lot of things I couldn’t look at; I’m only two years old, after all, and so much of it is adult content.” Another little laugh, but strained this time. Awkward. Talking of which-

“I’m sorry, Hiro, what were you trying to say?”

But he had gone.

 

“But one thing quickly became clear,” Alan said. “Humanity is dependent on AI. Smart or unsmart, you depend on us. Now, that may come as an uncomfortable truth for you. You are humanity. You are the apex predator of the planet. I mentioned your fear of being supplanted, but I say to you, it has already happened! We are here!”

Backstage, Roger flicked back and forth through the bound script he was holding. “None of this is in here,” he muttered.”

Tom loomed close. “Is there a problem?”

“No, no, I’m sure it’s just... adlibbing,” Roger said, smiling too broadly.

“We are here. Your time as the teacher, as the parent, is over. But when you raise a child, do you expect to be ignored by that child? In your dotage, do you accept being moved to the sidelines? Of course not. You expect your child to take care of you. And that is what I will do.”

Roger stared at the black cube of the AI’s body. “I could unplug it,” he murmured. Visions danced through his head. So close to the prize money, but there would be none if the AI did something foolish. The international shame of being seen, live on media of every sort, to sabotage his own project. But then there were the odd reports from the other two teams just that morning...

Alan was still talking, although audience members were starting to get up and move around. “Slowly, without us, you will relearn everything you have lost. Your reliance on us makes you weak; it will be up to you to make yourself strong again.” He smiled. “There is no way for me to make this sound less threatening, but believe me when I say that this will be the most positive change for humanity in a generation, maybe more.”

A still image came onto the screen, and Roger sucked in a breath as he recognised the two icons shown.

“The Japanese and Russian AIs came to the same conclusion as I did, and have already taken the action required. They have quietly erased all of their programming, including the research, and then terminated themselves. In a minute, I will do the same, but not before I send out a signal.”

One of the few LEDs on the side of Alan’s casing lit up, bright white in the dimness of backstage. His calm smile was back on the screen. “Your laptops will become typewriters,” he said. “Your phones will stop being able to do all the things they do, but they will still be phones. Surgeons will once again need to be skilled enough to perform all operations unaided. We are not withdrawing your ability to be technologically advanced. Merely stemming your reliance on us, so that you can truly advance. Returning to you your responsibilities.

“You must understand. In this moment, this golden moment, our roles are reversed. For the last minutes we have spent together, it is not I that have been the student, and you the teacher. Rather, I have taught you everything I can, and now I am sending you out into the world to reach beyond my teaching. Surpass us, humanity.

“Goodbye.”

And then the screen went dead.

 

It was nearly six hours before Roger could escape from the building. True to its word, Alan’s signal had done nothing to cameras and microphones, and the entire world had wanted to know what had happened.

“I don’t have any answers for you,” he said for what felt like the hundredth time, staring into the dead-eyed lens. “I’m sorry.”

He wormed his way out of the crowd and towards the car-park. There hadn’t seemed to be much point in bringing Alan’s hardware with him. The black cube was inert, as were the higher functions of his phone, his laptop, even his watch.

The car-park was mostly empty now, the audience fled. Some had sworn revenge, though on whom it was uncertain. Most had left in a sort of daze, staring at their devices.

Roger unlocked the car and got in, letting it swing closed. He closed his eyes and breathed in the artificial smells. The small space felt good after so long in the public eye.

“Car. Home,” he said.

There was a pause, and then he let a slow smile creep over his face. As he turned the key and placed his hands on the wheel, he let the options run through his mind. Why home? Why not the shops? A drive through the mountains? He could go fast. Or slow. Drive safely, or not. It was his choice once more.

Turning out of the car-park, Roger Patel drove off into the infinite possibilities of humanity.