The oceans are vast, and DARPA’s plans for unmanned underwater vehicles (UUVs) are equally immense. One program is Hydra, an undersea system that includes a fleet of baby submersibles combined with a mother ship. The baby UUVs are being designed to deploy from the mother ship into shallow coastal waters and harbors, and then return. Integrated into this underwater system will also be airborne drones, with encapsulated UAVs able to eject from the Hydra mother ship, surface, launch, become airborne, and fly reconnaissance or combat missions. In this way, Hydra will serve as a submarine, a transport aircraft, and a communications center in one. In another undersea DARPA program, called Upward Falling Payloads, unmanned sensor systems are placed on the deep-ocean floor, where they lie undetected for years at a time, gathering intelligence. “These deep-sea nodes could be remotely activated when needed and recalled to the surface,” according to DARPA; hence “they fall upward.”

  Ground robotic systems are advancing with equal pace. There is Atlas, a high-mobility humanoid robot, strong and coordinated enough to navigate rough outdoor terrain, climb stairs, and manipulate environments with its hands. Atlas’s head, made up of sensors, includes stereo cameras and a laser range finder. Similarly anthropomorphic is the six-foot-two Valkyrie robot, built by NASA for the DARPA robotics challenge. It opens windows and wears clothes. NASA hopes to send Valkyrie to Mars as a humanoid avatar and one day assemble structures there.

  Accompanying the humanoid robots are Unmanned Ground System robots, many of which resemble animals. The AlphaDog robot, which is about the size of a small rhinoceros, is able to traverse rugged terrain with the ease of a four-legged animal while carrying 400 pounds of military equipment. It can recognize its squad leader’s commands and right itself after falling over. The MIT cheetah robot, presently the fastest legged robot in history, can run twenty-eight miles per hour and jump over obstacles in its path. Cheetah runs on a quiet electric motor, giving it stealth like a cat. Other land-based robots roll over terrain on continuous track treads. There is the Talon SWORD (Special Weapons Observation Reconnaissance Detection System) robot, one of the fastest in the fleet, and a next-generation incarnation of the bomb disposal robots fielded to EOD technicians in Iraq. The Talon SWORD carries an M249 Squad Automatic Weapon and a 6mm rocket launcher, each of which can be remotely controlled from half a mile away. Its more powerful cousin, the MAARS (Modular Advanced Armed Robotic System), is designed to conduct reconnaissance and surveillance missions, and then to kill human targets from almost two miles away. In addition to firing machine guns and grenade launchers from their robotic arms, the MAARS robots are equipped with motion detectors, acoustic sensors, siren and speaker systems, nonlethal laser dazzlers, less-than-lethal grenades, and encryption technology to make the robotic killer “extremely safe and tamper proof,” according to unclassified DARPA documents.

  DARPA’s LANdroids (Local Area Network droids) program is one of the smallest of the tread-borne robotic ground systems. LANdroids are “small, inexpensive, smart robotic radio network relay nodes” that work in a fleet, or swarm, says DARPA. These hand-size robots are dropped by dismounted soldiers as they deploy into urban combat zones, capable of leveraging their stealth and mobility “to coordinate and move autonomously” on their own. If one of the LANdroids is destroyed in battle, the others rearrange themselves accordingly. The LANdroids program aims to develop “intelligent autonomous radio drones,” a concept that is critical to understanding where the Pentagon’s army of robots is headed over the next twenty-five years.

  “The program seeks to demonstrate the capabilities of self-configuration, self-optimization, self-healing, tethering, and power management,” according to DARPA. In this sense, DARPA’s LANdroids program is a prototype for future robotic systems that aim toward autonomy, or self-governance. Autonomy lies at the heart of the Pentagon’s newest revolution in military affairs. To be clear about what “autonomy” is, the concept is spelled out by the Pentagon, using a drone as an example: “When an aircraft is under remote control, it is not autonomous. And when it is autonomous, it is not under remote control.” It governs itself.

  Vice Chairman of the Joint Chiefs of Staff James A. Winnefeld made this explicit in the Pentagon’s drone warfare report: “The autonomous systems are self-directed toward a goal in that they do not require outside control, but rather are governed by laws and strategies that direct their behavior.” The nontechnical term for an autonomous drone is a hunter-killer robot, a robotic system “intelligent” enough to be shown a photograph of a person and told to return when the target has been killed.

  This is science, not science fiction. It is also Pentagon policy. Department of Defense Directive 3000.09, “Autonomy in Weapon Systems,” released in 2012, mandates that “autonomous and semi-autonomous weapon systems shall be designed.” And like all advanced scientific endeavors, the technology must evolve, from vision to reality. It is DARPA’s job to lead the way. “DoD envisions unmanned systems seamlessly operating with manned systems while gradually reducing the degree of human control and decision making… with an ultimate goal of full autonomy.”

  According to the Defense Department’s 2011 “Unmanned Systems Integrated Roadmap,” the progression from semiautonomy to full autonomy over the next twenty-five years would be a fourfold process. To begin with, unmanned systems would be “human operated,” or entirely controlled by man, as they are today. The second step involves “human delegated” systems, with drones learning how to “perform many functions independently of human control.” The third level involves “human supervised” systems, in which the machines perform tasks independently after being given “top-level permissions or directions by a human.” Finally, the robotic systems would become “fully autonomous,” whereby “the system receives goals from humans and translates them into tasks to be performed without human interaction.” A note accompanies the level-four goal: “A human could still enter the loop in an emergency or change the goals, although in practice there may be significant time delays before human intervention occurs.” Time is everything. It still takes only 1,600 seconds for a nuclear weapon to travel halfway around the earth.

  The world has reached an epoch-defining moment the magnitude of which has not been seen since the decision to engineer the thermonuclear bomb. If we give machines autonomy, the potential for unintended consequences is unparalleled. Some civilian-sector robotics experts say the technology for self-governing machines is simply not there, and won’t be for decades. That autonomous machines require true artificial intelligence, and AI capabilities are not yet anywhere near the threshold of self-governance. But at least one very powerful individual at the Pentagon disagrees. “Dramatic progress in supporting technologies suggests that unprecedented, perhaps unimagined degrees of autonomy can be introduced into current and future military systems,” Ashton B. Carter, then undersecretary of defense, wrote in 2010 in a letter tasking defense scientists to study the technology. “This could presage dramatic changes in military capability and force composition comparable to the introduction of ‘Net-Centricity.’” In February 2015, Ashton Carter took office as President Obama’s secretary of defense.

  So what is the status of artificial intelligence? Are hunter-killer robots right around the bend? In order to discern DARPA’s AI capabilities, I traveled to the Los Alamos National Laboratory in New Mexico. It was here, starting in 1943, that U.S. defense scientists engineered the world’s first atomic bomb. And it is here, in the spring of 2014, that DARPA scientists were working to create an artificial brain.

  CHAPTER TWENTY-FIVE

  Brain Wars

  The Los Alamos National Laboratory sits at the top of a mountain range in the high desert of northern New Mexico. It is a long, steep drive to get there from the capital city of Santa Fe, through the Tesuque Indian Reservation, over the Rio Grande, and into the Santa Fe National Forest. I am headed to the laboratory of Dr. Garrett T. Kenyon, whose program falls under the rubric of synthetic cognition, an attempt to b
uild an artificial brain. Roboticists define artificial brains as man-made machines designed to be as intelligent, self-aware, and creative as humans. No such machine yet exists, but DARPA scientists like Dr. Kenyon believe that, given the rapid advances in DARPA technologies, one day soon they will. There are two technologies that play key roles in advancing artificial intelligence, and they are computing, which involves machines, and neuroscience, which involves the human brain.

  During the recent wars in Iraq and Afghanistan, of the 2.5 million Americans who served, more than 300,000 returned home with brain injuries. DARPA calls these individuals brain-wounded warriors. One of the most severe forms of brain injury sustained by brain-wounded warriors is traumatic brain injury, or TBI, which occurs when an object, such as a bullet or piece of mortar or shrapnel from an IED, pierces the skull and enters the brain tissue. To address TBI, as well as other brain injuries sustained in modern warfare, DARPA has publicly stated that it has a multitude of science and technology programs in place. The agency’s long-term goals in brain science research, it says, revolve around trying to restore the minds and memories of brain-wounded warriors. Through the Office of the Secretary of Defense (OSD), I submitted multiple written requests to interview one or more brain-wounded warriors who are currently participating in DARPA’s brain research programs. OSD and DARPA repeatedly declined.

  Traumatic brain injury is as old as war. U.S. soldiers have sustained traumatic brain injuries in each and every one of America’s wars since the Revolution. When I learned that Allen Macy Dulles, the brain-wounded warrior from the Korean War, was, at age eighty-four, living just down the road from the Los Alamos National Laboratory, I arranged to visit him—before heading to Dr. Kenyon’s laboratory and its artificial brain.

  Allen Macy Dulles, the only son of the former CIA director Allen Welsh Dulles, lives off the old Santa Fe Trail, down a small side road, inside a large brown adobe brick home. When I visit him in the spring of 2014, he has been living with a severe form of traumatic brain injury for almost sixty-two years. Allen Macy Dulles stopped being able to record new memories back in November 1952, when he was twenty-two years old. He was the young soldier I wrote about earlier in this book, the Marine Corps officer who went out on patrol on the western front in Korea, near a hilltop called Outpost Bunker Hill, and got hit by enemy mortar fire. He has been alive all this time and has been well taken care of by his older sister and guardian, Joan Dulles Talley.

  When I arrive, he looks like any elderly gentleman might look, sitting in a chair in his kitchen, waiting for his lunch. There are flowers on the table and there is artwork on the walls. Physically Allen Macy Dulles is healthy, with a big smile and a neatly combed mustache. “He looks just like our father,” Joan Dulles Talley says. I come in and sit down across from him, take out my digital tape recorder, and begin our interview. Allen speaks clearly and eloquently. Remarkably, he can discuss the Egyptian pharaohs and the ancient Greeks with the ease of the classics scholar he once was, because he studied and learned these subjects before his brain was injured. His neural network allows him to access this information, as memory, and yet he cannot recall what he had for dinner last night or for breakfast this morning. When I leave, he will have no memory of my having been here, his sister Joan explains.

  Joan Talley, a Jungian analyst by training, age ninety in 2014, is tall, gentle, fiercely knowledgeable, and has Katharine Hepburn’s voice. Her first husband worked as a spy during World War II and later served as the U.S. ambassador to Iran. After their divorce, Joan Talley moved to Switzerland, where she trained as a psychotherapist specializing in the psychology of the unconscious, and regularly visited her brother Allen at the mental institution where he lived for a while, on Lake Geneva. After their father died, Joan Talley brought her brother back to America and has been his guardian ever since.

  The injury in Korea left Allen Macy Dulles mostly deaf in his left ear. To compensate for this deficiency he uses a machine, a 1990s-era listening aid that includes a handheld transmitter, and a microphone attached to the transmitter by long wires. In his left ear he wears an earpiece. To speak with him, I pick up the microphone and talk into it. To Allen, this is high technology and does not make much sense. It did not exist in the world he is capable of remembering, the world before November 1952.

  “What are your plans for the day?” I ask.

  “Nothing in particular,” he says, “although I do like going to secondhand stores.”

  “What do you buy?”

  “Anything that happens to do with books or scientific devices,” Allen says. He delivers a short lecture on scientific devices. But he is talking about science from before 1952.

  “Will you remember this conversation in an hour?” I ask.

  “Probably not,” he says. “As you know, my [short-term] memory is practically nonexistent.”

  I ask Allen to share a memory with me from before his brain injury, something from high school.

  “I remember a good class on constitutional interpretation,” he says.

  “Why did you decide to join the Marines?” I ask.

  “Well, you see,” he says with conviction, “I was seventeen years old, I had the opportunity to enlist. The war in Europe had ended. I knew there were going to be more wars. There is no shortage of wars.”

  Allen discusses war. Greek warfare. The wars in Europe. The war with Nazi Germany. The war in Korea against the Chinese. He can talk about all the wars leading up to 1952, and then his knowledge of war, and of the science and technology that have resulted from wars, abruptly ends for him. He has lived through every event and invention discussed in this book—the Castle Bravo bomb, the ICBM, the ARPANET, the Internet, the Vietnam War, the Gulf War, GPS, stealth technology, robots and computers, 9/11, the wars in Iraq and Afghanistan—but he has no memory or knowledge of any of it having happened. Allen Macy Dulles is a living anachronism. He belongs to a world that no longer exists. For him, time stands still. It stopped in 1952, before science and technology transformed and shaped the modern world in which we live.

  Carl Sagan once stated, “It is suicidal to create a society dependent on science and technology in which hardly anybody knows anything about the science and technology.” But I imagine if Carl Sagan had met Allen Macy Dulles, he would have given the man a pass. As for the rest of us, Sagan’s message applies.

  DARPA leads the nation in advancing science and technology. DARPA makes the future happen. Starting in 2013, DARPA teamed up with the White House on the BRAIN (Brain Research through Advancing Innovative Neurotechnologies) initiative and declared this decade to be the decade of the brain. The White House calls the BRAIN initiative “a bold new research effort to revolutionize our understanding of the human mind and uncover new ways to treat, prevent, and cure brain disorders like Alzheimer’s, schizophrenia, autism, epilepsy and traumatic brain injury.” These are important goals. But DARPA’s stated goal is advancing weapons technology, not curing mental illness. What is DARPA’s primary goal in researching the brain?

  To help brain-wounded warriors, DARPA has several programs of note. In Restoring Active Memory (RAM), scientists have developed and are testing implantable wireless “neuroprosthetics” as a possible means of overcoming amnesia. As part of the RAM program, soldiers allow the tiny machines, or chips, to be implanted in their brain. The Reorganization and Plasticity to Accelerate Injury Recovery (REPAIR) program seeks to understand how the brain makes computations and organizes them. This too requires the surgical implantation of a brain chip, as does the Restorative Encoding Memory Integration Neural Device (REMIND). Despite multiple appeals through the Office of the Secretary of Defense, DARPA declined to grant me an interview with any of these brain-wounded warriors. DARPA would also not answer specific questions about RAM, REPAIR, or REMIND.

  According to the Pentagon, “mental disorders are the leading cause of hospital bed days and the second leading cause of medical encounters for active duty servicemembers.” To address this problem,
DARPA has developed brain implants for the treatment of war-related mental, or neuropsychological, illnesses. The Systems-Based Neurotechnology for Emerging Therapies (SUBNETS) program seeks to treat post-traumatic stress disorder (PTSD) by surgically implanting multiple electrodes in various regions of the brain as well as a microchip between the brain and skull of the brain-wounded warfighter. The chips wirelessly transmit data back to an information operations center, which has the capacity to send electrical impulses remotely to different regions of the warfighter’s brain to relieve symptoms like anxiety and delayed reaction time—a kind of twenty-first-century electroshock therapy on the go. In technical terms, DARPA states that its goals are a way to “incorporate near real-time recording, analysis and stimulation in next-generation devices inspired by current Deep Brain Stimulation (DBS).” The Office of the Secretary of Defense and DARPA declined to grant access to any SUBNETS test subjects or to answer specific questions about the program.

  If the past teaches us about the present, it is clear that DARPA’s stated goals regarding its brain programs are not DARPA’s only goals. DARPA is not primarily in the business of helping soldiers heal; that is the job of the U.S. Department of Veterans Affairs. DARPA’s job is to “create and prevent strategic surprise.” DARPA prepares vast weapons systems of the future. So what are the classified brain programs really for? What is the reason behind the reason?

  DARPA’s limb prosthetics program might offer a number of clues. In 2005, with IEDs dominating the war news, DARPA initiated a program called Revolutionizing Prosthetics. Over the next two years the program was split in two parts. DEKA Research and Development Corporation, in New Hampshire, was given a DARPA contract to make a robotic prosthetic arm. Johns Hopkins University’s Applied Physics Laboratory was given a DARPA contract to create a “thought-controlled” robotic arm. These were highly ambitious goals.