The robot revolution is a long-standing staple of science fiction, from Metropolis to The Terminator, and the spread of advanced machines in recent years has done little to contradict its seeming inevitability. But before master and servant switch roles, the robots of the world need to conquer a simple task: efficiently recognizing a cat. In 2014, it took some 16,000 computers running on a “neural network” set up by Google to accomplish this feat—something a toddler does by reflex. The algorithms for determining “catness” are rapidly improving, but this distinction between thinking and calculating is a great divide that robotics engineers will need to breach before robots can truly replace humans. A kid knows that all things with four legs, such as chairs, aren’t cats. IBM’s Watson supercomputer triumphed on the quiz show Jeopardy in 2011, but it also thought Toronto was in the United States.
Humans And Robots, Mutually Beneficial
Robots are increasingly moving into more corners of the world. One study estimated that as many as 45% of jobs in the United States would be eligible for robot replacement over the next 20 years. Law offices have already reduced the demand for paralegals because computer programs can now process legal briefs and synthesize large documents. People have worried about being replaced by machines since the start of the industrial age. Although this has proved to be a legitimate concern in many jobs, machines can also be a complement to human endeavor rather than a substitute for it. Machines and computers, in general, have amplified human strength through mechanized tools, augmented human intelligence through information processing, and extended human reach through telecommunications and remote sensing.
But these robots couldn’t operate autonomously. They needed instruction—a person to think up the process and rules for each task and audit the outcome—just to be sure that the shipment from Amazon.com has a book about cats in the box, not one about woodworking or parakeets. This online retailer among the major adopter of robots, and a major advocate for collaboration between humans and machines. At Amazon’s massive product warehouses, fleets of heavy-lifting robots bring shelves of products to a packaging employee, so the employee doesn’t have to spend time locating the correct product. According to a recent report, the company now has 200,000 robots working in its warehouse.
Going Where No Human Can Go
Hospitals rely on robots for tasks that range from hauling laundry to assisting with delicate surgeries. The world’s first complete robot surgery took place in 2010 at McGill University Health Center in Montreal, Canada. A pair of robots called DaVinci (the surgeon) and McSleepy (an anesthesia bot) removed a patient’s prostate. A surgical team controlled DaVinci’s arms via video control while observers in the room monitored the operation closely. Robotics allowed for greeted steadiness and precision in handling the instruments than a human could achieve, but the operator still needs to decide where and when to cut or stitch—at least for now.
Thousands of military and police robots have been deployed successfully to do work that is too dangerous for humans. Robots can clear away land mines in war zones, inspect suspicious packages on city streets, and provide 3-D imaging of unstable buildings. Real-life robots can scout potentially hostile environments and keep soldiers safe, like the PackBot made by iRobot, which is used to identify explosives, chemical weapons, and radioactive materials.
In March 2011, a tsunami devastated parts of Japan and severely damaged reactors at the Fukushima Daiichi nuclear plants. The radiation released resulted in the evacuation of 300,000 people and the development of a clean plan that will span decades. In that unsafe environment, humans turned to robots for damage analysis, radiation monitoring, and debris removal.
Researchers in Australia have built a microscopic robot that can move around in a manner similar to E. coli bacteria. The tiny machine could take a biopsy from inside the human body.
What Is A Robot?
What distinguishes a robot from other technologies or machines is not always clear-cut, in part because scientists have not yet come to a consensus on which machines qualify as robots. Generally, robots are autonomous or semi-autonomous that can manipulate the physical environment through their moving parts or suggestions. At a basic level, robots need only to sense something and put what they are sensing through a decision algorithm and act. They may resemble humans or other living things, and their range of capabilities may mimic “thought” and autonomous fiction.
Humanoid devices with a broad array of abilities and decision-making skills are becoming more common, from Honda’s Asimo, which resembles a person in a spacesuit, to Baxter, a robot factory worker. Rethink Robotics, the company that created Baxter, says that one of its advantages is trainability—you can show it what to do rather than program instructions. But today’s robots still depend heavily on people, from robotic arms and other factory equipment that needs the correct coordinates to know where to go to the many military devices that need orders from a human operator. However, as artificial intelligence research advances, robots may someday be capable of making their own choices.