Robots Rising

Worldwide, there are many developments in robotics, several of which are weirder and wackier than a talking version of a dead novelist and a robot board member.

In May this year, Hong Kong became the base of Hanson Robotics, a company that was founded in the US to develop and manufactures bio-inspired technologies for humanlike robots, complete with naturalistic rubber skin, and neuro-inspired artificial intelligence. Its website features a video with a robot version of science writer Philip K. Dick, telling a smiling interviewer, “Don’t worry, even if I evolve into Terminator I will still be nice to you.”

            While the wires protruding from the back of the head show the Dick robot is far from becoming akin to the Terminators of movie fame, it is part of a fast evolving, wide ranging field. Elsewhere in Hong Kong, May also saw venture capital fund Deep Knowledge appoint a robot to its board. A robot of sorts, that is: rather than moving, the new board member is a machine learning program capable of making investment recommendations.

            Worldwide, there are many more developments in robotics, several of which are even weirder and wackier than a talking version of a dead novelist and a robot board member. Some are comical, a few are downright scary.

            Perhaps the scariest is South Korea’s SGR-1 robot sentry. Developed by Samsung Techwin, this stands around half the height of a human yet may be more than a match for any gunslinger. There’s no rubber skin, no attempt to resemble a person. Instead, it totes a machine guns plus a grenade launcher, which can be swivelled and aimed based on input from heat and motion sensors capable of detecting humans up to two miles away.

            Though a promotional video recalls scenes from Robocop, there’s no concern that SGR-1 might become self aware and rampage against its creators, as it requires a remote human operator to give permission to open fire. Since 2006, an undisclosed number of these sentries now keep watch along the border between North and South Korea. They’re never lazy, never lose concentration, and are always watching.

            To you and me, the robot sentry might seem downright sinister. Yet Samsung Techwin spokesman Huh Kwang-hak was quoted in the Daily Mail as saying, “The SGR-1 can and will prevent wars”.

            I guess Kwang-hak would see potential for even greater “war prevention” in a robot that researchers at MIT unveiled this week. Dubbed a robotic cheetah, this is roughly the size of a family dog, and can run at up to 10 miles per hour as well as leap over foot-high hurdles. Though still slow compared to humans, the researchers think they can triple its speed, making it more than a match for sprinter Usain Bolt. Perhaps we’ll soon see a dystopian movie featuring a platoon of machine-gun equipped cheetah robots, rampaging across the land to enforce order according to the whims of their programmers.

            Happily, such scenes are still restricted to science fiction. Yet last month came a glimpse of how multiple robots can act in concert. A Harvard University team showed off a swarm of tiny robots, each a few centimetres across, which can organise themselves to form two-dimensional shapes like a star or the letter K. With each robot only able to communicate with and measure distances from each of its neighbours, the robots collaborate with one another until the overall pattern is created. Noting similarities with biological systems, lead researcher Radhika Nagpal, Fred Kavli Professor of Computer Science at the Harvard School of Engineering and Applied Sciences, said, “At some level, you no longer even see the individuals; you just see the collective as an entity to itself.”

            Robots are also working with people. Collaborative robots now roam some factories, performing a variety of tasks that were too onerous for humans, who might be free to perform other duties – or free to seek work elsewhere. These are smaller and cheaper than the large robots that hitherto featured on production lines, and instead of being caged for safety reasons, they have sensors so they can slow down or halt if a collision appears likely. One of the best known collaborative robots is called Baxter, and even has a computer screen to signal “emotions” such as puzzlement.

            While if you would someday like help with the housework, take note of Boris. This is a new British robot that can assess unfamiliar objects, and determine how best to pick them up with a five-fingered hand on a robotic arm. Though billed as being one day able to load a dishwasher, Boris copes well with cups and plates, while struggling with knives and forks.

            So far, all robots have had to be taught or learn for themselves about such objects. That could change, with a project using cloud technology. Aiming to create a wealth of knowledge in the cloud, for access by robots, the project is called RoboBrain. If you’ve seen any of the Terminator films, you may immediately think of similarities with Skynet, the artificial intelligence that controlled defence systems and one day unleashed them against the human race. For now, though, RoboBrain is far more basic, and you can visit the website to find it has recently learned information such as how to grasp an umbrella or a mouse, and the appearance of a heatmap for a seated human who is reading a book.

            With developments snowballing, scientists including renowned physicist Stephen Hawking recently warned of the potential perils of artificial intelligence. In an article for the Independent newspaper, they wrote that there are no fundamental limits can be achieved, adding, “So, facing possible futures of incalculable benefits and risks, the experts are surely doing everything possible to ensure the best outcome, right? Wrong.” Little serious research is being conducted regarding the pitfalls; yet now is the time to ask how to improve the chances of reaping the benefits and avoiding the risks.

            While risks from killer robots loom large in the popular imagination, it is even possible that some of us humans will become like robots’ sex slaves. Robot girlfriends have already appeared, albeit at least one simply lies still, though talks when touched. Today, Professor Adrian Cheok of City University London is giving the keynote talk during the Festival of the Mind, on “telepresence” technologies which will blend the virtual and the physical and may even result in sexual and emotional relationships between humans and robots.

            As you ponder where such things may lead, I’m more concerned about research finding robots can concoct passable news reports. Yikes! Could this mean my byline will someday vanish, replaced by the tireless, utterly objective science writer, RoboDoc?

Martin Williams