Forget about Wall Street, ignore the analysts, blot out all acronyms containing a “2,” and take a walk into our robotic future at the Massachusetts Institute of Technology Artificial Intelligence Lab.
There’s Kismet with its big eyes, slightly curved lips, and quirky animal ears. Sit in front of it, and you’ll be amazed by how it engages in natural, almost human, interaction. It smiles, scowls, and then turns demure and coy. You may find yourself following suit if you get a chance to meet it. And you will — or at least in the future, you’re very likely to meet a robot that’s modeled after Kismet and its conversational technology.
Down the hall from Kismet is Cog, a tall and lanky collection of metal, wires, and microprocessors that vaguely resembles a person with its humanlike torso and arms. Cog stands idle and disinterested at the moment. But a team of researchers is hard at work on a common-sense idea that is really rather revolutionary.
Babies learn by mimicking, so why can’t robots? Sounds logical, so MIT researchers are programming Cog to watch human behavior through its video-camera eyes, and then mimic, remember, and reenact movements and mannerisms just as humans do. Here, the possibilities are boundless.
“Just What Do You Think You’re Doing, Dave?”
Robots may still seem more like science fiction than science, but in the past several years, robot technology and artificial intelligence have progressed far beyond R2-D2 and the Grand Master. And while much of the world is preoccupied with deciphering the Internet and its revenue streams, big thinkers have moved on to the robotic future.
Thinkers as diverse as Rodney Brooks, director of the MIT AI Lab; Ray Kurzweil, an author and inventor; and Marcia McNutt, president and CEO of the Monterey Bay Aquarium Research Institute, believe that robots will become as commonplace as home computers in the not-too-distant future. Perhaps the idea of the Jetsons is not so absurd after all.
In his most recent book, The Age of Spiritual Machines: When Computers Exceed Human Intelligence (Viking, 1999), Kurzweil predicts that by the late 2020s, computer chips — and thus, robots — will possess the computing power of the human brain. Whether that will put robots on par with humans is up for debate. But just to stir the controversy, Kurzweil further contends that man will become more robotic as he adds technology to his body in the form of memory chips and retinal implants. Kurzweil also asks the question, Will humans ever accept robots as conscious beings?
Brooks — who is also the cofounder, chairman, and chief technical officer of iRobot, a company that makes and sells robots — imagines a future where robots will benefit and expedite existing businesses, rather than simply blow our minds with metaphysical questions. He believes that today’s robotic technology is only as advanced as computer technology was in 1978. Travel back to that future, Brooks says, and you can predict robot technology tomorrow.
There Goes the Neighborhood
As early as the 1950s, computers existed within big companies. The same is true about robots today (think welding robots in auto plants). Before computers became consumer products, hobbyists started making them from do-it-yourself kits and began tinkering with toys that ran on tiny computer chips. Today, computers are ubiquitous. Brooks and his compatriots expect robots to follow the same lifeline — only much, much faster.
Of course, Brooks acknowledges that scientists need to solve some big problems before robots begin to act like humans. “Robots today aren’t very good at pattern recognition — understanding whether something or someone is inanimate or animate, old or young, male or female,” he says. “We need several Einsteins to deal with those issues.”
To understand the future of robots — and the reality of that future — people will need to expand their definition of what a robot is and does. Science fiction has contributed no shortage of cool, sometimes scary notions of life on earth with robots — how they’ll look, what they’ll do, when they’ll suck our brains dry. Much of that will remain fiction forever.
Where No Man Has Gone Before
The AI lab and its industrial partners expect to introduce robotic attachments like knees and legs for humans long before it launches a full-fledged, humanlike robot like Data from Star Trek. Indeed, business and medicine may be the first industries to root robots in reality.
Though robots made for business purposes wouldn’t necessarily look or act like humans, they would be able to “think”: to analyze data and determine their own courses of action. For instance, the ocean may soon contain robotic fish that can take advantage of their environment.
Take McNutt’s work at the Monterey Bay Aquarium Research Institute. McNutt envisions a future where ocean exploration no longer means launching big, expensive voyages; it’s instead made up of ongoing explorations “manned” by robots. “Expeditions are good for understanding what lies in our solar system,” she says. “But we don’t know how our ocean changes over time because we can’t set an expedition out for years at a time.”
One idea is to build submerged stations filled with robots that gather and analyze data. That information could be sent to the ocean’s surface, where it could then be transmitted intermittently. (It’s difficult to transmit data wirelessly through seawater.) “I need something that can go deeper, stay longer, and won’t be power-hungry,” McNutt says. “And it has to be smart on its own, without human intervention.”
McNutt believes that robotic sea explorations will take place by about 2025, but no one knows for sure when the technology will be ready. “Thinking about the future gives scientists something to dream about,” she says. Brooks remembers a time when he was asked to describe the world in 3000. “That’s like asking people in the year 1000 what they thought today would be like,” he says. “They probably would have predicted sharper writing quills and colored inks.”