Sorry, science fiction fans, but the “replicants” of the Blade Runner saga or the “terminators” of the eponymous action movie franchise are not on the horizon.
“Don’t imagine human-like, humanoid robots when you think of the future of robots,” said Kim Sang-bae, the world-renowned robot scientist who developed a four-legged walking robot called “Cheetah,” which has gained widespread media exposure.
Not only is it impossible to develop human-like robots now, it may remain impossible in the future, according to Kim, a mechanical engineering professor at the Massachusetts Institute of Technology.
While the ultimate stage of robotics may, indeed, be creating machines that can think and work on their own, there is a yawning gap between where robot technologies stand now and that final-stage development.
In an interview with Asia Times, Kim predicted that the robot industry would continue to expand by creating robots which can do very specific things better than humans. But he conceded there is a real risk that the social inequalities in the sector will accelerate.
“It’s hard for human-like robots to appear,” Kim said. “People tend to confuse computing power and human intelligence, but just a little digging shows how different the two are.”
Specificity versus generality is one key difference. “Robots that drive like, or better than humans, can appear with lightning-quick processing of data such as roads and maps.”
This kind of task-specific programming, however, should not be compared to the wide spectrum of human intelligence. “Computers can do complex calculations very well,” he said. “They can play Go, or drive like humans, but these abilities are completely different from human intelligence.”
Despite some speculative reports and alarmist predictions, AI, or artificial intelligence, has real limits in replicating human intelligence, which encompasses emotions and the ability to reason and analyze.
“There is no technology that can make machines understand concepts and judge values like humans – there is just no concept about how to implement this,” Kim said. “Many people expect human-like humanoid robots to be created, as in science fiction [but] it’s hard even to imagine that at the moment.”
Challenges
Moreover, the challenges of creating human-like robots extend beyond AI to engineering. Even a natural hand gesture – a base-level physical task for humans – is “technologically challenging to implement in robots,” Kim said. And even if that engineering challenge is surmounted, commercialization looks unviable.
Kim estimated that a single humanoid robot for household work would be extremely expensive. “We can make a high-performance washing machine or vacuum cleaner for a much lower price with a few motors and valves!” Kim said. “Humanoid robots are hard to establish as business models.”
So, where does the future of robotics lie? Apparently, not too far from where we already are, with specialized, completely non-humanoid machines – for example, robots that assemble cars or reconnaissance and predator drones that are changing the face of warfare.
“Robots will eventually develop in a way that they function efficiently in certain areas rather than in humanoid robots,” said Kim, although he conceded that certain robots could, feasibly, take humanoid shapes if those shapes are fit for particular purposes.
Kim’s current research is in robot mobility. He specializes in applying animal movements to robots – a concept that birthed his most famed work, “Cheetah.” His research is ultimately aimed at creating a robot that can go anywhere and move freely, specifically to operate in disaster zones and rescue scenarios.
Along with these high-mobility machines, Kim is also developing high-sensitivity remote-control devices that allow the person controlling the robot to sense the forces and obstacles that the robot encounters.
He firmly believes that robots should work for people. In addition to his work on robots that can operate in dangerous environments, he is also interested in machines that can care for the elderly in aging societies – a challenge currently facing both Korea and Japan.
Robot risks
These realities are far from humanoid robots of science fiction, which, programmed with artificial intelligence, move and operate and act on their own initiatives. He is not concerned about robots becoming the masters and humans the slaves. “It’s unrealistic to worry that robots like Terminator will appear and threaten the survival of mankind,” he said.
Yet, he concedes that robots do, in fact, pose real risks for mankind. “When we talk about the coexistence of robots with people, we should think about how to maintain social structures fairly, not be mired in unrealistic agonies,” he said.
While robotics may assist some businesses – such as factories or taxi companies – to increase their earning, robots also replace human jobs by merging low cost with high efficiency.
“Just as the distribution of wealth was exacerbated after the industrial revolution, the widening use of robots may bring about a similar kind of problem,” Kim said. “Fair distribution, rather than worries about Terminator-like robots, is what we need to consider when we think of the peaceful coexistence of robots and people.”