Robots have long captured the public’s imagination, as their blend of technology and seeming humanness leads to intriguing questions, excitement, and concern
In the mid-20th century, most people asked would probably imagine that, by 2018, we’d all have a robotic assistant in our homes and that most employees would be working alongside somewhat humanoid robotic counterparts. In reality, we’ve only recently seen some of the anticipated advancements in robotics come online. As modern robots become more and more capable, it’s important to understand the opportunities, questions, and potential threats robots pose.
Robotics is essential in modern manufacturing. However, industrial robots typically stay in one place, and they generally don’t look much like humans at all. Robots excel in assembly-line tasks, and it’s likely they’ll continue having an impact in manufacturing. Modern robotic components have improved in terms of precision, and better sensors enable robots to perform complex tasks that traditionally require humans. However, these types of robots are still inherently limited and expensive to develop. The goal for many researchers is robots that excel at more general tasks.
Sensors and Sensibility
Not long ago, digital cameras used to be prohibitively expensive. Today, however, even the cheapest smartphones on the market typically have two cameras built in, and the resolution these tiny cameras produce is impressive. Furthermore, other types of visual sensors, including lidar, can provide better precision and 3D sensing capabilities. As large companies, including heavyweights such as Google, Honda, and Boston Dynamics continue investing in robotics, their ability to interact with the world in a safe and sophisticated manner will improve. It’s perhaps no coincidence that today’s advances in robotics come at a time when the Internet of Things, itself based heavily on sensors, is thriving.
Traditionally, robots are portrayed as hard and mechanical, a reputation partially due to the types of robots used in manufacturing. This can make robots seem unapproachable and even dangerous, as small mistakes can lead to significant injuries. An emphasis on soft robotics is changing these perceptions and will prepare robots to thrive in medicine and other fields. The public in general is somewhat wary of robots. Being able to hug one and have it hug back might defray some of these concerns.
The term “killer robots” is one the tech field would like to stay away from, but the potential of robots in warfare is immense. Simply put, it’s fairly easy to attach a few actuators to a basic robot and program it to fire a weapon, and a small fleet of armed robots can wreak havoc on soldiers and civilians alike. Twenty-six nations have called for an outright ban on fully autonomous robots, but the ease with which remotely controlled robots can be converted into autonomous ones might mean these developments are inevitable. We’re generally happy to have robots build items we use or, perhaps, deliver our meals, but there are grave concerns about robots tasked with deciding who lives and who dies.
One of the concerns about robots in warzones is based on the reality experts have largely accepted: The software powering robots is still based on old technology. In the 1950s and 1960s, it seemed inevitable that software would be able to replicate human-like thinking. Today’s artificial intelligence, however, is still based on older concepts and doesn’t seem to be moving, in a qualitative manner, toward human-like reasoning and creativity. The concept of neural networks, which power many of today’s most advanced AI systems, dates back to the 1940s, and sophisticated approaches, including backpropagation, dates to the mid-1970s. Although we can program increasingly sophisticated AI systems, technology is still limited, and expectations about advances in robotics need to take this into account.
In the 1970s and 1980s, Western societies were concerned about robotics eliminating all manufacturing jobs. The incursion of robots into the workforce, however, proved to be a slow process, and Western countries moved more toward non-manufacturing fields. However, robotics is now threatening the massive service industry, and it seems inevitable that jobs will be lost as robots become more sophisticated. Societies will need to deal with the aftermath of this move. If the number of people a restaurant employs drops significantly, where will displaced workers go? If labor can be automated, how will this added wealth be distributed?
When companies aim to create robots that look like humans, they can often come remarkably close. However, nobody has yet been able to craft a robot that acts convincingly human for more than a few seconds. The uncanny valley phenomenon, where robots look eerily close to humans, raises a question: What do we want robots to look like?
— CNBC (@CNBC) May 27, 2018
Should a robotic assistant be designed to look like a human, or is a robot that looks more like Wall-E or Honda’s Asimo more appropriate? Replicating human speech is a remarkable achievement, but perhaps we’re better served with robotic voices that sound distinctly robotic, as blurring the line between humans and robots might have undesirable effects.
Perhaps the most pressing issue with robotics is determining what regulation is necessary. Recent incidents involving partially self-driving cars has brought the issue to light. While in development, robotic systems will almost inevitably fail. But what safeguards exist to ensure companies are acting responsibly, and do governments need to step in to create guidelines? The first company to develop a safe self-driving car stands to benefit tremendously, and this potential financial benefits can encourage negligent behavior. As robots become more capable, society at large will need to address how to encourage innovation in a safe and ethical manner.
From self-driving cars to person robotic chefs to in-home assistants, robotics holds tremendous promise for changing society. However, robotics has also lead to drone warfare, and other types of remotely operated robots have lead to concerns about a future where warfare is cheap and too easy to justify. It may be some time before we all have a personal robot assistant to help with day-to-day tasks, but there’s little doubt that we’ll be seeing robots more and more frequently.