by Joel Garreau
One thrust ends up making machines more lifelike. “Neurotechnology for biomimetic robots,” he says. “Getting robots to jump, run, crawl, do things that nature does well. We’re evolving our machines to be more like animals.”
The other thrust connects life more directly with machines. “Let’s create higher-density interconnects with living systems. And let’s do so with the brain and with neural tissue.” What does this mean? Rudolph describes it as “listening to the orchestra of neurons.”
“Much of what you and I do that’s different than a cockroach is based on our central nervous system and our brain,” he observes. “It was just a leap of faith that if we could create interfaces with living systems”—devices that can “listen to the plethora of signals, that good things would happen.” If you listen to the “orchestra” of the brain, you might be able to detect patterns in the signals “and try to make a song or a piece of music that had meaning.”
The result is massive connections between individual living neurons inside the skulls of humans and wires that lead to computers. The first commercial step was cochlear implants—tiny machines that allow the profoundly deaf to hear by wiring tiny computers directly to the nervous system. “You talk about transforming humans. That was one of the earliest examples of a successful brain-machine interface. Thirty-five thousand people now have cochlear implants and are doing pretty well,” Rudolph observes. The next step is retinal implants—computer eyes—wired to the brain of the blind.
But that’s not the hard part. The DARPA-hard part is hearing the symphony for motion. “That’s a really hard defense problem,” Rudolph says.
“One of the fun things is you go play combat zoologist,” he says. “You take a bunch of scientists and you go stick ’em in an infantry unit and you run around and play the war game. You go down to Combat Town in Quantico. You go on a carrier and get a ride in an F-16. It’s a real fantasy camp. It’s a great way to get a sense of the problems. We can be future thinkers, but when you have a customer you can’t lose sight of that. And as long as we have a D in front of DARPA, we have to be cognizant of the customer’s interest.
“I’m sure people raise their eyes and say, ‘What the hell is a brain-machine interface going to do for the Defense Department?’ Well, you draw them back to Foxfire with Clint Eastwood. The cockpit of a jet fighter is a complex place to be, filled with challenges for human performance. If the brain can somehow play a role in command and control of those systems, that would be a good thing.”
Here’s where Gina Goldblatt comes in.
“We’ve got a program in exoskeletons here. A guy’s going to put on a suit and run and jump high or whatever.” This is the Superman program in which the suit picks up on its wearer’s muscle movements and greatly amplifies them so, for example, you can carry a 500-pound pack on your back for a very long time. But suppose the exoskeleton was responding not to your muscle movements but to your brain commands. “The reality is if you had a brain-machine interface, that’s going to be an integral part of how he does that well. We’re asking a fundamental question: Can the brain accommodate control and command of new devices?”
That was the point of Belle controlling the mechanical arm with her thoughts. Or of Ivy and Aurora, the monkeys that came after Belle. Eventually you might fly an F-22 with your thoughts. Or Gina might control artificial muscles with her mind.
Right now the way you make it happen with monkeys is not great. You drill a hole in the head of the monkey and implant a device that looks like a very tiny hairbrush bristling with hundreds of wires, each of which lines up next to individual neurons. “We are thinking now of ways to get these interfaces in without opening your head, drilling the thing in,” says Rudolph.
“We’ve got our team of 70 crazed academics charging this dream. There are commercialization efforts sort of working now in parallel. A new company started called Cyberkinetics. They’re going to sell a device to quadriplegics, paraplegics and locked-in patients who can’t move a damn thing. Christopher Reeve had intact central nervous system motor function. So you can implant a chip in those people.”
Then comes the question of how you run a device that you’re not used to running. “I’ve got a little robotic dog and a little robotic cockroach. The cool thing is can I figure out how to control and experience that from a [human] central nervous system point of view?”
Isn’t the hard part of that getting the information from the robotic cockroach and porting it back into your skull?
“Yes, the same chip will close the loop and will allow you to experience—whatever—controlling a robot, maybe flying a plane. Not from the standpoint of feeling it as a joystick, but as other sensory input. Visual, mechanical, force dynamics.”
In other words, you should be able to pipe any sense from any sensor, anywhere, into your brain. You might directly sense the images from a remote camera, for example, allowing you to feel as if you had eyes in the back of your head. For that matter, you might feel a color or taste a sound.
“That’s the powerful vision that I think this could enable. There’s no reason why one can’t input with a chip other types of experiences.”
Closing the loop—allowing the human brain to receive signals directly, not just send them—opens the door to genuine telepathy. Rudolph’s researchers are working on creating telepathic marmosets. Marmosets are very small South American monkeys with thick, soft, richly variegated fur. He’s trying to get them to conduct brain-to-brain communication.
“Marmosets have distinctive calls associated with fear and threat, food and familial identification. We’re going to use that device—that hairbrush. Both of them are going to have hairbrushes” embedded in their brains. “We’re going to send the pattern with that call to a second one. If he hears the right call, we’re going to look for a response. Then the question is, well, what if the monkey says, ‘Fuck you’—you know what I mean? There are a number of issues and challenges that we’re going to face with this. A National Academy member is involved with this one, at Vanderbilt—Jon Kaas, the guy who mapped the motor cortex. I mean, I couldn’t have been more pleased to get a senior guy willing to take that kind of chance. It is amazing what real critical thinkers are willing to do.”
Why stop at transmitting speech directly into the brain? Why not pictures?
“Can I alter what you see, change what you see, or put something in that you see what I see or you see what my camera sees?” Rudolph says. For example, “I want to see over the hill and I send a micro air vehicle or a robot over there and now I’m experiencing the visual image of the robot. I see what it sees.”
There’s no reason, by the way, that these images have to be in the spectrum of ordinary light. If you want to see in infrared or ultraviolet or whatever else the machine can sense, patch those night-vision puppies right into your visual cortex. Why should owls have all the fun?
“The third project,” says Rudolph, “has profound implications for neuro degeneracies as well as augmented humans. I think it’s the closest in terms of thinking about the evolution of man in the context of cognition. You’re going to put the hairbrush in with this chip that mimics the circuitry of the brain.” You replace a damaged portion of the brain with a chip that works like the brain. “For the first time you have a chip that now participates with brain function. Such a chip could be used to augment brain function. You start to increase processing speed. You will enhance memory. We’ll know in the next year or two whether we can replace a lost circuit.”
How are you going to change the batteries?
“It’s a good point. It’s a good point. Power is a big issue. Our battery technology sucks. Our power problems are huge. I think all these implants will be run off the energy in the body, ATP. There’s low-temperature fuel in the body. The body is amazing in terms of its chemical conversions of energy. So we have a whole program that we launched.”
Is that yours, too?
“Biomotors, yes. I share it with another pro
gram manager, Anantha Krishnan. We have some implantable batteries that work off of the natural body constituents. Tissue engineering is going to give us muscle. Building robots with living muscle.”
Why would you want to do that?
“Right now we can keep it alive longer than we can get a battery to work. Yes. Outside the body. Yes. We’ve got a thing called the ‘lox bot.’ ” It’s a little biorobotic device that resembles a piece of smoked salmon. “It uses skeletal muscle from a frog, and the damn thing swims using skeletal muscle. It swims through its energy source. It’s in a bath of glucose and ATP and the thing swims for like 20 hours. That’s the University of Michigan and MIT.”
The challenge, I tell Rudolph, is going to be convincing my mother that this is not science fiction.
“Me too.” He laughs. “You’re just trying to write about it. Try telling them I’m spending your tax dollars doing this.
“And we haven’t talked about bees. I do a lot of stuff with bees. That’s 10,000 flying dogs. They can be trained to sniff for things. They are little electrostatic dust mops so they collect things on their body and bring back spores and all kinds of information. A honeybee hive of 10,000 or so makes 100,000 trips per day over a five-kilometer radius. So the amount of information coming back to the hive is huge. So we place technology at the hive and just monitor the hive. We looked at one where you pull in a truck with concealed explosives and train honeybees to smell explosives. Put the honeybee hives on either side and with a camera just look at the truck and count bees swarming around the truck.”
Does former UN weapons inspector Hans Blix know about this?
“Oh, you know, the military is way too conservative to use a beehive. We told this to the Israeli national police, and they grilled me for, you know, 45 minutes on what happens if somebody gets stung. So I said, ‘Look, at the end of the day you’ve got a choice. You can get stung by a bee or blown up by a terrorist. I’ll take the bee.’”
Have you read Michael Crichton’s book, Prey?
“I haven’t read his book yet. Is it good?”
Well, the bad guys are funded by DARPA.
“Oh, Jeez,” he replies.
The list of DARPA-inspired human enhancements goes on and on. General Dynamics has a development and production contract potentially worth $3 billion intended to transform muddy-boot soldiers into nodes on a network. This involves manufacturing what the Pentagon calls “uniforms” for use by soldiers in the field by 2010. But this Objective Force Warrior Ensemble is far more than clothing. The soldier wears an undershirt fitted with body sensors that keep track of and broadcast his vital signs. His helmet receives video from robots. It also holds a camera, night-vision amplifiers, infrared sensors, laser finders, a global positioning system and a skull-mounted transmitter and receiver. A retractable eyepiece is useful for reading text messages or to view images sent from command centers or drones. Body armor is lighter, contributing to an equipment weight reduction of 50 percent, to 50 pounds. In this configuration, unmanned vehicles, known as mules, carry supplies. “Can you imagine traversing the mountains of Afghanistan with 100 pounds on your back?” asks program engineer Jean-Louis “Dutch” DeGay.
The original vision of the Engineered Tissue Constructs (ETC) program is based on the idea of rebuilding customized organs and body parts on demand, with the construction going on inside your body, not transplanted.
One of the goals of the Metabolic Engineering program is to allow badly injured soldiers to go into suspended animation or hibernation. It would allow them to survive even without oxygen for short periods of time, until the area is safe enough for help to arrive. This is also the program interested in allowing soldiers to run Olympic-quality sprints for 15 minutes on one breath of air. Turns out humans are very inefficient in the way we process resources. There’s a whole lot of oxygen in one breath, and we waste most of it.
The Bioinspired Dynamic Robotics program is trying to replicate the foot of the gecko, the tropical lizard with amazing feet that perform Spider-Man–like feats. It would be handy for robots also to be able to climb straight up walls and hang from ceilings.
The Mesoscopic Integrated Conformal Electronics (MICE) program has already succeeded in printing electronic circuits on the frames of eyeglasses and helmets, weaving them into clothes, even putting them on insects. These include electronics, antennas, fuel cells, batteries and solar cells.
The Biological Input/Output Systems program is designed to enable plants, microbes and small animals to serve as “remote sentinels for reporting the presence of chemical or biological” particles. They’d do this by changing color, lighting up fluorescently, dropping their leaves or changing the color of their flowers.
The Brain-Machine Interface program is investigating how you would put wireless modems into people’s skulls.
And that’s just the Defense Sciences Office, the department of DARPA most directly involved with human enhancement. Meanwhile, on the floor where the Information Processing Technology Office (IPTO) resides, its director, Ron Brachman, former research vice president at AT&T Labs and previously at Bell Labs, and president of the American Association for Artificial Intelligence, wants to complete DARPA’s vision from the sixties. When the original IPTO was created in 1962, its director, J.C.R. Licklider, focused the office on his novel conception of computers and humans working in symbiosis. That idea resulted in the Internet. Now the new IPTO “wants to realize this vision by giving computing systems unprecedented abilities to reason, to learn, to explain, to accept advice, and to reflect, in order to finally create systems able to cope robustly with unforeseen circumstances,” according to Brachman. The object of the game is to produce machines—and the italics are his—“that truly know what they’re doing.”
Some of this is so far-out-sounding that it beggars description. Don’t even ask about the “special focus area” called Time Reversal Methods, for example.
Devotees of the film Men in Black may recall the scene near the end when the two protagonists sit wiping their faces of intestinal slime from the interstellar cockroach they have just vanquished. Will Smith turns to Tommy Lee Jones and says, “This definitely rates about a 9.0 on my weird-shit-o-meter.”
Many is the time, cruising DARPA, that it is easy to recall that scene.
Readers with eclectic historical memory may by now be asking, “Aren’t these the same guys who, during the Cold War, poured our taxpayer dollars into crackpot schemes like extrasensory perception and remote viewing?” Yes, and the guys at DARPA still don’t apologize for it. In fact, it’s a perverse badge of honor. “If those had worked, wouldn’t you like to have known about it?” asks Goldblatt. “As long as it can be investigated rigorously and systematically and step by step, very little is too far out for us.”
Readers with an interest in civil liberties, meanwhile, may by now be asking “Aren’t these the same guys who created that Total Information Awareness project with the crazy logo of the Egyptian pyramid and this all-seeing eye, the implication of which was that in the name of anti-terrorism, we would never again have the slightest shred of privacy?” Wasn’t that the one run by former admiral John Poindexter of Iran-contra fame, who later was canned from DARPA in a flap over a proposed futures market in terrorism? Actually, Poindexter worked not in the Defense Sciences Office but in a different department. Whole other floor. And they’ve canned the logo and changed the name. And they’re feeling hurt and misunderstood. The press jumped to all sorts of inaccurate conclusions, they say. Not surprising given that DARPA, as usual, displayed its maddening reflex of not wanting to discuss what’s up, even when the outlines of their projects are on the public record. But yes, that’s DARPA.
Will all these projects work? Unquestionably, no. Not all. DARPA’s attitude is that if an idea looks like a sure thing, let somebody else fund it. The National Science Foundation. Or venture capitalists, more to the point. A project is regarded as “DARPA-esque” only if few others would tackle it, but it would
be earth-jolting if it did work. If you don’t have failures, you’re not far enough out. DARPA managers view themselves as instigators. By the time something new is mainstream enough to attract academic conferences attended by several hundred researchers, DSO usually sees its midwife work as done and moves on to new challenges. At the same time, DSO ruthlessly cuts off money to projects that fail to achieve their milestones, goals and objectives. An effective program manager knows when to cut bait.
Will all these projects bear fruit soon? Some more than others, and for the same reasons. The bulk of DARPA’s projects operate in the 5- to 10-year time frame. But especially in the Defense Sciences Office, by the time a project is sufficiently mature that manufacturers are asking what color seats you’d like in it, program managers, like the Lone Rangers they are, have disappeared in a cloud of dust and a mighty “Hi-yo Silver.” At the same time, history is moving fast enough that portions of this list may very well become part of your life between the time this is written and the time you read it.
Do defense dollars cause weird bounces in the research efforts? No question. The program to grow replacement organs from scratch, for example, as exciting and promising as it seemed, didn’t initially do much for the Pentagon. “That’s Veterans Affairs,” they said.
What this recitation does demonstrate, however, is that engineered evolution of humanity is in the works right now. This inventory is hardly theory, much less fiction. This is about real flesh-and-blood human beings doing substantive, material things. It is a snapshot of what one small portion of one organization is working on in the first decade of the 21st century. It does not begin to include everything else coming out of labs and institutes around the world. Yet it shows that researchers are hardly waiting for some vague future. They are not, for example, waiting for some technology such as gene modification to mature in the next decade. They are working on enhancing people in important ways right now.