One day, Lipson asked two of his PhD students to bring a demo chatbot to his Artificial Intelligence class. Acting a bit like a portable, rudimentary psychotherapist, a chatbot is an online program that reflects what someone says in slightly different words and asks open-ended questions. It can come across as surprisingly lifelike (which says a lot about the clichés that pass for everyday chitchat). But in 1997 a “Cleverbot,” designed by the British AI expert Rollo Carpenter, went online with a teeming arcade of phrases compiled from all of its past conversations. Each encounter had taught it more about how to interact with humans, including the subtleties of innuendo and pricks of friendly debate, and it learned to apply those nuances in the next chat. Since then it’s held twenty million conversations, and its verbal larder is a treasury (or a snakepit) of useful topics, ripe phrases, witty responses, probing questions, defensive expressions, and the subtle rules of engagement, gleaned from years of bantering with humans.
Lipson’s grad students set the laptops face-to-face on a table so that they could tête-à-tête in a virtual parlor. On one screen a computer-generated male materialized, on the other screen a female. The man spoke with a slight British accent, the woman in a syncopated Indian voice. Fortunately, the grad students videotaped the encounter and posted it online, where the chatty Cleverbots have now enchanted over four million people with their oddly human conversation.
The robots begin with a simple “Hello there,” followed by pleasantries, but as they respond to one another they soon start to disagree, and the exchange grows funny, poignant, snarky, and thoroughly hypnotic.
“You were mistaken,” Mr. Cleverbot says to Ms. Cleverbot, adding sarcastically, “which is odd, since memory shouldn’t be a problem for you!”
“What is God to you?” she asks him at one point.
“Not everything,” he says. It’s a surprisingly plausible answer.
“Not everything could still be something,” she insists with jesuitical aplomb.
“Very true,” he concedes.
“I would like to believe it is.”
“Do you believe in God?” he asks.
“Yes I do,” she says emphatically.
“So you’re Christian . . .”
“No I am not!” she snaps.
They bicker and make nice-nice. He calls her a “meanie,” for not being helpful. She suddenly asks him a painful question, one any human might wonder about. Still, it’s disquieting to hear.
“Don’t you want to have a body?”
And then, surprisingly, like someone who has accepted a fate he nonetheless laments, he answers: “Sure.”
What else is there to say? Abruptly they freeze into replica humans once more, and the video clip is over. Some people detect animosity or sexual tension between the man and woman, others a marital spat. We’re ready to accept fictional robots in movies and stories, but are we ready for a synthetic life form that feels regret, introspects, and conducts relationships—creatures opaque to us, whose minds we can’t fully mirror? Do the chatbots appeal because they’re so like us, or because we’re so like them?
There are scores of people in robotics who can fine-tune a robot’s movements, even design truly lifelike robots with delicately mobile faces. Italian roboticists, for example, have created a series of realistic-looking heads that synchronize thirty-two motors hidden beneath the robots’ polymer skin, and mimic all of our facial expressions, based on muscle movements, and can even capture the emotional space between furrowing the brows, say, and frowning. Such robots have already passed the stage of being a mere sensation in the robotics world. Fully-featured human faces are smiling, grimacing, exchanging knowing looks the world over. Unlike Madame Tussaud’s wax-museum stars, today’s robots look lifelike enough to seem a bit creepy, with facial expressions that actually elicit empathy and make your mirror neurons quiver. Equally realistic squishy bodies aren’t far behind. One can easily imagine the day, famously foretold in the movies Blade Runner and Alien, when computers with faces feel silicon flavors of paranoia, love, melancholy, anger, and the other stirrings of our carbon hearts. Then the already lively debate about whether machines are conscious will really heat up. This was always the next step toward designing a self-aware, agile, reasoning, feeling, moody other, who may look like you or your sibling (but have better manners).
No doubt “robot sociology” and “robot psychology” will emerge as important disciplines, because there’s an interesting thing that happens when robots become self-aware. Just like people, they sometimes get wrong impressions of themselves, skewed enough to create robot delinquents, and we might start to see traits parallel to psychological problems in humans.
When I used to volunteer as a telephone Crisis Line counselor, it wasn’t always easy finding ways to help the callers who phoned in deep despair or creased by severe personality disorders. Self-aware robots with social crises, neuroses, even psychoses? That might prove a challenge. Would they identify with and prefer speaking to others of their kind? Suppose it concerned a relationship with humans? Colleges have popular schools of “International Labor Relations,” “Human Ecology,” and “Social Work.” Can “Interspecies Labor Relations,” “Robot Ecology,” and “Silicon Social Work” be far behind? How about a relief order for aged, infirm, or incarcerated robots, such as “Android Daughters of Charity” or “Our Sisters of Perpetual Motion?”
What would the Umwelt (worldview encompassing thoughts, feelings, and sensations) of a self-aware robot be like? We’re no longer entertaining such ideas merely as flights of imagination, but contemplating how to behave in a rapidly approaching future with the startling technology we’re generating. If, as Lipson says, our new species of conscious, intelligent robots will learn through curiosity and experience, much as children do, then even robo-tots will need good parenting. Who will set those codes of behavior—individuals or society as a whole?
CAN WE LIVE inside a house that’s a robotic butler, protector, and chatbot companion all rolled into one, an entity with its own personality and metabolism? Its brain would be a robotic Jeeves (or maybe Leaves), who tends the meadow walls and human family with equal pride, and is a good listener, with a bevy of facial expressions. A fully butlered house with a face that rises from a plastic wall would monitor the energy grid, fuel the car (with hydrogen), while exchanging news, ordering groceries, piloting a personal drone to the post office, and preparing a Moosewood Restaurant lunch recipe that includes herbs from the herb-garden island in the kitchen, and arugula and tomatoes from the rooftop garden. In some high-tech enclaves, smart locks are now opened by virtual keys on iPhones, and family members wear a computer tracking chip that stores their preferences. As they move through each room, lights turn on ahead of them and fade away behind, a thermostat adjusts itself, the song or TV show or movie they were enjoying greets them, favorite food and drink are proffered. The house’s nervous system is what’s known as the “Internet of Things.”
In 1999, the technology pioneer Kevin Ashton coined the term for a cognitive web that unites a mob of physical and virtual digital devices—furnace, lights, water, computers, garage door, oven, etc.—with the physical world, much as cells in the body communicate to coordinate actions. As they cabal among themselves, synchronizing their energy use and activities, they can also share data with the neighborhood, city, and wired world.
Combining animal, vegetable, mineral, and machine, his idea is playing out in the avant-garde new city of Songdo, South Korea, where the Internet of Things is nearly ubiquitous. Smart homes, shops, and office buildings stream data continuously to a cadre of computers that sense, scrutinize, and make decisions, monitoring and piloting the whole synchronous city, mainly without human help. They’re able to analyze picayune details and make sure all the infrastructure hums smoothly, changing traffic flow during rush hour as needed, watering parks and market gardens, or promptly removing garbage (which is sucked down through subterranean warrens to a processing center where it’s sorted, deodor
ized, and recycled). Toiling invisibly in the background, the council of computers can organize massive subway repairs, or send you a personal cell phone alert if your bus is running late.
It’s a little odd thinking of computers taking meetings on the fly and gabbing together in an alien argot. But naming it the Internet of Things domesticates an idea that might otherwise frighten us. We know and enjoy the Internet, already older than many of its users, and familiar now as a pet. In an age where even orangutans Skype on iPads, what could be more humdrum than the all-purpose, nondescript word “things”? The Internet of Things reassures us that this isn’t a revolutionary idea—though, in truth, it is—just an everyday technology linked to something vague and harmless sounding. It doesn’t suggest brachiating from one reality to another; it just expands the idea of last century’s cozy new technology, and animates the idea of home.
In J. G. Ballard’s sci-fi short story “The Thousand Dreams of Stellavista,” there are psycho-sensitive houses that can be driven to hysteria by their owners’ neuroses. Picture sentient walls sweating with anxiety, a staircase keening when an occupant dies, roof seams fraying from a mild sense of neglect. Some days I swear I’m living in that house right now.
PRINTING A ROCKING
HORSE ON MARS
For centuries, the world’s manufacturing has been a subtractive art, in which we created artifacts by cutting, drilling, chiseling, chopping, scraping, carving. As a technology, it’s been both mind-blowing and life-changing, launching the Industrial Revolution, spawning the rise of great cities, spreading the market for farm-raised goods, and wowing us with everything from ballpoint pens to moonwalkers. It’s still a wildly useful method, if sloppy; it creates heaps of waste and leftovers, which means extracting even more raw materials from the earth. Also, mass-produced items, whether clothing or electronics, require a predicament of cheap labor to add the final touches.
In contrast, there’s “additive manufacturing,” also known as 3D printing, a new way of making objects in which a special printer, given the digital blueprint for a physical item, can produce it in three dimensions. Solidly, in precise detail, many times, and with minimal overhead. The stuff of Star Trek “replicators” or wish-granting genies.
3D printing doesn’t cut or remove anything. Following an electronic blueprint as if it were a musical score, a nozzle glides back and forth over a platform, depositing one microscopic drop after another in a molten fugue, layer upon layer until the desired object rises like a sphinx from the sands of disbelief. Aluminum, nylon, plastic, chocolate, carbon nanotubes, soot, polyester—the raw material doesn’t matter, provided it’s fluid, powder, or paste.
Hobbyists share their favorite digital blueprints via the Internet, and some designs are licensed by private companies. Like many other technologies, 3D printing does have a potential dark side. People have already printed out handguns, brass knuckles, and skeleton keys that can open most police handcuffs. Future laws will undoubtedly restrict access to illegal and patented blueprints, and also to dangerous metals and gases, explosives, weapons, and maybe the fixings for street drugs.
Imagine being able to press the print button whenever you want a candelabra, toothbrush, matching spoon, necklace, dog toy, keyboard, bike helmet, engagement ring, car rack, hostess gift, stealth aircraft rivets—or whatever else need or whim dictates. The Obama administration announced that it had seen the future and was investing $1 billion in 3D printing “to help revitalize American manufacturing.” According to scientists and financial analysts alike, within a decade household 3D printers will be as common as TVs, microwaves, and laptops. However, people will still need to buy supplies and copyrighted blueprints for home printing, and many will order 3D objects ready-made from cottage industries.
In the future, even in the Mars colony Olivine calls home, she could fabricate a rocking horse of exactly the right height and dappled pattern on the morning of her daughter’s birthday. Or she might print an urgently needed pump, and then a set of demitasse spoons with Art Deco stems. Or paint shades that don’t yet exist in tubes. Artifacts that can’t be created in any other way, such as a ball within a ball within a ball within a ball. Or an item with a hundred moving parts that’s printed as a single piece. From this strange new forge, who knows what artworks and breakthroughs will emerge. The creative opportunities are legion.
We may ignore all the traditional limits set by conventional manufacturing. With micrometer-scale precision, we can seal materials within materials, and weave them into stuff with bizarre new structural behaviors, like substances that expand laterally when you pull them longitudinally. A brave new world of objects.
What is an object if you can grow it in your living room drop by drop or molten coil upon coil? How will we value it? Today, because 3D printing is still a novelty for many people, we value its products highly, in wonderment. But when cheap home 3D printers become commonplace (today’s cost anywhere from $400 to $10,000), and factory 3D printing replaces the assembly lines and warehouses, and even body parts and organs can be made to order, we’ll live in an even more improbable world, where some objects continue to exist as tangible things, as merchandise, but a great many will exist concretely but in nonmaterial form, in a cloud or in a cartridge of fluid or powder, the way e-books do, as quickly accessible potential.
As cars, rockets, furniture, food, medicine, musical instruments, and much more become readily printable (some of those already are), it’s bound to temporarily unnerve the world’s economies. After all, we value things according to their scarcity. When gold is plentiful, it’s cheaper. But if objects lurk as software codes, inside computers, and are abundantly available at the push of a button, they’ll exist as another class of being. How will that change our idea of matter and the physical reality of all that surrounds us? Will it lead to an even more wasteful world? Will handcrafted objects become all the dearer? Will the Buddhist doctrine of nonattachment to worldly things flourish? Will we become more reckless?
This may all seem far-fetched, but not so long ago the Xerox machine was a leap of faith from carbon paper. When I first worked as a professor, making a carbon copy—what the “cc” on e-mail stands for—was a part of daily life. It’s still somewhat astonishing to me that we can now print images in color, from home machines that can connect to our computers through the air.
Many companies won’t look the same, because they won’t need to hire scores of workers, buy raw materials, ship or stock or produce anything. Industry, as we know it, may end. Financial advisers, business magazines, and online investment sites such as the Motley Fool believe 3D printing companies will clean up big-time, because their overhead will be so much lower, and they’ll sell only the clever designs or raw materials.
Not right away. Most people will probably still find it more convenient to buy ready-made things. But soon enough, in the next fifteen years, 3D printing will revolutionize life from manufacturing to art, and practical visionaries like Lipson feel certain it will usher in the next great cultural and psychological revolution. For some, that future is the obvious sequel to the digital revolution. For others, it’s as magical as a picture painted on water.
“Just like the Industrial Revolution, the assembly line, the advent of the internet and the Social Media phenomenon,” Forbes magazine forecasts, “3D Printing will be a game changer.”
How close are we to that day? It’s already dawned. 3D printers are whipping up such diverse marvels as drone aircraft, designer chocolates, and the parts to build a moon outpost from lunar soil. Already, the TV host Jay Leno uses his personal 3D printer to mint hard-to-find parts for his collection of classic cars. The Smithsonian uses its 3D printer to build dinosaur bones. Cornell archaeologists used a 3D printer to reproduce ancient cuneiform tablets from Mesopotamia. Restorers at Harvard’s Semitic Museum used their 3D printer to fill in the gaps of a lion artifact that was smashed three thousand years ago. In China’s Forbidden City, researchers use a 3D printer to inexpensively restor
e damaged buildings and artworks. NASA used 3D printing to build a prototype of a two-man Space Exploration Vehicle (an oversized SUV astronauts can live in while they explore Mars). A USC professor, Behrokh Khoshnevis, has devised a method known as Contour Crafting for printing out an entire house, layer by layer—including the plumbing, wiring, and other infrastructure—in twenty hours. When 3D printers are linked to geological maps, houses can be made to fit their terrain perfectly. Khoshnevis is designing both single houses and colonies for urban planning, or for use after hurricanes, tornadoes, and other natural disasters when fully functional emergency houses will be 3D-printed from the ground up.
Boeing is 3D-printing seven hundred parts for its fleet of 747s; it’s already installed twenty thousand such parts on military aircraft. The military’s innovative design branch, DARPA, which began funding 3D printers two decades ago, finds them invaluable for repairing fighter jets in combat or supporting ground troops on the front lines. They’re superb at coining parts instantly, remotely, to exact specifications, without having to wait for urgently needed supplies, or risk lives to ferry them through hostile terrain. Companies like Mercedes, Honda, Audi, and Lockheed Martin have been fashioning prototypes and creating numerous parts inside 3D printers for years. Audi plans on selling its first 3D-printed cars (modules printed then robot-assembled) in 2015.
The Swiss architect Michael Hansmeyer has 3D-printed the world’s most complex architecture: nine-foot-tall Doric columns of breathtakingly intricate swirling organic laces, crystals, grilles, pyramids, webs, beehives, and ornaments, madly rippling around, fainting through, vaulting from, and imbedded into each other as layers of exquisitely organized chaos that began as a mirage in the mind and hardened. Containing sixteen million individual facets and weighing a ton, it looks like a roller-coaster ride down a scanning electron microscope into the crystalline spikes of amino acids. It’s easy to imagine a cathedral by Antoni Gaudí with such columns in Barcelona. Or the labyrinthine short stories the Argentine fabulist Jorge Luis Borges might unleash among them.