TOP TREND: MID-YEAR UPDATE
2016 FORECAST: Robots will continue to take on human roles, both large and small, and for better and worse. Robotic cars park themselves, sparing us the anxiety of manual parallel parking on a busy street, but also robbing us of a skill; drones — flying robots — spy on our enemies, but also enable your neighbor to spy on you. In 2016, a robotic fighter jet will be ready to relieve human pilots confronting overwhelmingly complex situations.
Among the most ubiquitous robots in the near future will be those used as personal assistants. These robots, often taking human-like shapes and wearing something that looks like a face, can provide companionship to homebound elders, monitoring their moods and medication use and offering rudimentary forms of conversation. At the other end of the age spectrum, child-like robots are coaxing children with autism out of their emotional isolation.
Surgical robots, which have tendrils outfitted with tiny drills and micro-scalpels, are being guided by human surgeons using joysticks to carry out complicated procedures where human hands are too big to go. After proving themselves on bone, in 2016 surgical robots will begin testing their skills on soft tissue.
The most significant aspect of robots’ rise is their marriage to artificial intelligence. Instead of needing to be programmed with explicit instructions, robots can learn through trial and error — a process known as “machine learning” or “deep learning” — and also can transfer their knowledge to each other. In Ashutosh Saxena’s lab at Stanford University, for example, a robot learned to boil eggs by watching internet videos, then “taught” a faraway robot to do the same by sending its knowledge over the web. Robots’ ability to learn will be a greater growth area than the evolution of their physical design.
While robots are already being used to stack warehouse shelves, scan and check out store items and drive certain types of vehicles, the next wave of robotics will assume middle-management responsibilities. These include quality control, data collection and analysis and resolving customer complaints.
Various researchers are refining robots’ physical appearance and capabilities, creating more human-like faces and sensor-laden robot “skin” that can detect, for example, that a robot is touching something on fire. Intense development also continues around robots’ emotional affect, enabling machines to interact more effectively with humans.
British technopreneur Patrick Levy Rosenthal claims to have embedded 12 emotions on a chip that can be installed in robots to give them human-like feelings.
But the greatest advances continue to be made in robots’ ability to access knowledge and operate independently. A consortium of researchers is creating “RoboBrain,” a giant repository of basic information, videos, appliance operating manuals and other practical wisdom that any robot will be able to access. For example, if a robot has never seen a coffee mug, it can tap into RoboBrain to learn what a mug is for, that it can be held by the handle and that it has to be carried upright when full but not when empty. Engineers at Tufts University have given a robot the ability to reflect on an instruction and refuse to carry it out if the robot thinks the action might endanger it. TJ