He presses the button on the podium, and the satellite photo on the screen is replaced by an image of software code. Hundreds of lines of instructions, written in a programming language I don’t recognize, run from the top of the screen to the bottom. “This is a portion of Sigma’s source code. When we developed the software for the AI, we focused on imitating human skills such as reasoning, language, and pattern recognition. We succeeded in creating a self-aware intelligence that could accomplish almost any task a human can perform, from proving a mathematical theorem to composing an opera. But in one important respect, Sigma was a failure. We weren’t able to give it humanlike morality or motives. Sigma has no incentive to pursue what’s good for the human race because it lacks the ability to empathize.”
Dad presses the button again, and this time a photo of chimpanzees comes on the screen. “Empathy comes naturally to humans because it played a big role in our evolution. The most successful apes were the ones who could imitate and understand each other. Sigma, in contrast, has no empathy. It’s aware of our presence, of course, and it even sent a couple of messages to our military headquarters, but the AI has blocked all our attempts to communicate with it. The basic problem is that Sigma’s intelligence is very different from ours. We don’t understand the AI, and it doesn’t understand us either. So we need to build a bridge between us and the machine.”
He pauses, as if to gather his courage. Then he presses the button once more and a diagram of the human brain appears behind him. Just below the familiar organ is a close-up view of a section of brain tissue, magnified to show the individual brain cells and the many branchlike connections between them. Clinging to the cells are hundreds of tiny golden spheres. They look like bits of pollen.
Dad steps toward the screen and points at the spheres. “These are nanoprobes. Each is less than a thousandth of a millimeter wide. We can make trillions of them in the lab.” He reaches into the pocket of his uniform and pulls out a vial of yellowish fluid. “In fact, I have several trillion probes right here, floating in this liquid. If we inject enough of these nanoprobes into a human brain, they’ll spread throughout the organ and stick to the brain cells. If we then scan the brain with X-ray pulses, the probes will absorb the energy and start to glow. The scanner will record the positions of the glowing dots attached to the cells, and their patterns will give us a detailed map of all the connections within the brain and the strength of those connections.”
His voice is getting louder. That often happens when Dad talks about his research. He can’t help it; he gets excited. “This is the key,” he says, holding the vial of nanoprobes up to the light. “All our memories, all our emotions, all our quirks and virtues and flaws—all that information is stored in the connections between our brain cells, which create new links or alter the old ones whenever we learn or remember something. So if we make a sufficiently detailed scan of a person’s brain, we’ll have a full description of his or her personality, which can be held in an electronic file of about a billion gigabytes. The next step is downloading that information into circuits that mimic the cells of the human brain. We already have that kind of neuromorphic circuitry because we built it to hold our AI software.”
The audience is murmuring again. Some people are confused. And some, like me, are terrified, because they can see where this is going. Shannon Gibbs leans forward and points at the screen. “Are you talking about making copies?” she asks. “Copies of our brains?”
“Yes, exactly. Once the information is downloaded into the neuromorphic electronics, the circuits will replicate the connections of the person’s brain, re-creating all its memories. And as data flows through the circuits, the electronic brain will generate new thoughts based on these memories. Just like in a human brain, the thoughts will organize themselves into a conscious intelligence, a self-aware entity that can set goals for itself and communicate with others, either by text or through a speech synthesizer. And the ‘personality’ of this new intelligence would be identical to the one inside the person’s head, because it would be based on the same memories and emotions and character traits.”
Shannon wrinkles her nose. She looks queasy. “Have you…tried doing this yet? Making a copy of someone?”
Dad nods. “Four months ago we tried the procedure on three volunteers. All were Army veterans with high IQs. Unfortunately, the experiment failed each time. We scanned their brains and successfully downloaded the data into the circuits, but in each case the human intelligence failed to run on the computer. We were able to copy their minds, but the copies didn’t survive the transfer.” He furrows his brow. “Since then we’ve studied the problem, and now we know what went wrong. The crucial factor is the person’s age. After the age of eighteen, there’s a change in the structure of brain cells. They become coated with greater amounts of a substance called myelin, which insulates the cells and makes them more rigid. This increases the efficiency of a person’s thinking but reduces its flexibility. The mind of an adult is simply too inflexible. It can’t adapt to the new conditions of residing in a machine.”
“So now you’re going to try to copy younger minds?”
He nods again. “We were planning to conduct the next phase of the experiment later this year, but the events in Russia have accelerated our plans. This time, all the volunteers must be sixteen or seventeen. At that age you’ve reached your maximum brainpower but your minds are still adaptable. In addition to being highly intelligent, the volunteers must have strong, resilient personalities.” Dad sweeps his arm in a wide arc, gesturing at all the teenagers in the room. “All of you meet those requirements.”
Shannon rears back in her seat as if she’s been slapped. “And where are you going to store the copies of our brains?” Her voice is furious. “In a supercomputer? A big electronic prison?”
Dad doesn’t take offense. He answers her calmly. “The scanning process converts human intelligence to a digital form, allowing it to run on any neuromorphic computer that has enough memory and processing power. But in the initial stage right after the transfer, we believe it’s important to connect the intelligence to a machine that can move around and sense the outside world. A human intelligence is accustomed to controlling a body, so if we want to preserve its sanity, we’d better give it something to control. Here, let me show you.”
He puts the vial of nanoprobes back in his pocket and pulls out something else, a small remote-control device. He points it at the doorway beside the stage, and a moment later I hear a loud clanking. The noise startles the soldiers standing by the doorway. They step backward, flattening themselves against the wall. Then a seven-foot-tall robot emerges from the doorway and brushes past them.
The robot strides across the stage. It has two arms and two legs, but otherwise it isn’t very humanlike. It has no head or neck. Its torso is shaped like a giant bullet, with the rounded end on top. Its legs angle downward from the base of its torso and rest on oval steel-plate footpads that clang against the floor.
The machine marches briskly past the podium and stops in front of my dad, who presses a button on his remote control. This command extends the robot’s arms, which telescope to a full length of six feet. They look like multi-jointed tentacles. The machine’s hands, though, resemble human hands, with dexterous mechanical fingers and thumbs.
Dad presses another button, and the robot’s rounded top starts to turn like a turret. “The cameras and acoustic sensors are up here,” Dad says, pointing at the top end. “But the neuromorphic electronics are deep inside the torso, encased in armor plating. These robots were originally designed for the war in Afghanistan, so they’re pretty sturdy.” He raps his knuckles against the torso. “All in all, it’s an excellent platform for a newly transferred intelligence, but really it’s just the beginning. The whole point of the Pioneer Project is to bridge the gap between man and machine, and that means the human intelligences must explore their new environment. The Pioneers will have to learn how to use their new capabilities, and that includes transferring their intelligences from one machine to another.”