Yes, the robots are coming, but they are not conquering the planet anytime soon — even if that’s what the media hype would have us believe. And they’re not as smart or technologically advanced as we think, according to Kate Darling, a leading expert on robot ethics.
“I’m not worried so much about robots developing their own agenda and taking over the world,” Darling told delegates at the 71st CFA Institute Annual Conference in Hong Kong. “I’m a little bit more worried about people and how we decide to use the robots as a society.”
Darling is a researcher at the Massachusetts Institute of Technology (MIT) Media Lab, where she investigates social robotics and conducts experimental studies on human-robot interaction. She says artificial intelligence (AI) needs “massive sets of data” to learn, and so there are a lot of incentives for companies to collect as much data as possible. At the same time, consumers don’t have much incentive to curb this trend, as data collection ties directly to the functionality of their devices. Think, for example, of Alexa, Amazon’s digital assistant, which is built into other Amazon devices.
“Over the long term, we are going to see privacy massively eroded because of this,” Darling said. “We can manipulate robots, but robots can also manipulate us, or the companies creating the robots can also manipulate us and a lot of these technologies I see being developed right now are being developed specifically for very vulnerable parts of the population, such as the elderly and children, and we might need to think about ways to protect them.”
“With any new technological development, there is lot of hype and a lot of fear that the robots are going to take over the world and kill us all,” she said.
“It’s important to anticipate problems before they happen and to try to find solutions well in advance,” Darling explained. “But, on the other hand, it’s a little frustrating because the media picks up on this as the one problem we need to be concerned about in this future of robotics, and I think that that distracts from some problems we might actually want to turn our attention to and leads people to overestimate where we are with the technology.”
She acknowledged that there are some areas where robots are much better than humans. “They can do math,” she said. “They can remember everything. They can work tirelessly on an assembly line. They can recognise patterns in data. They can beat us at Go and at Jeopardy.”
But, there are areas where they lag.
“If you ask a robot to understand context, or ask a robot to understand concept, or transfer skills from one context to another, or deal with anything unexpected that happens, the robots are woefully and hopelessly lost,” she said. “That’s not going to change anytime soon because we don’t even know where to start to begin to develop the type of artificial general intelligence that’s required to, for example, context switch the way that a person can.”
That all said, robots are becoming an increasingly common sight.
In Shanghai, the world’s first fully automated, human-free bank branch is open for business. Bank customers are greeted by Xiao Long, or “Little Dragon,” who speaks with them, accepts their bank cards, and answers simple queries.
“What I really love about the way that people interact with robots, is that we treat them a little bit like they are alive even though we know perfectly well that they are just machines,” Darling said. “Part of this is because we’re primed by science fiction and pop culture to want to personify these machines, and that is partly why we are constantly comparing them to humans”
She also believes humans are biologically conditioned to anthropomorphize robots, to want to see ourselves in them. “We generally have this tendency to want to project human-like qualities and emotion onto non-humans and it’s something we do from a very, very early age and we think we do it in order to make sense of non-human entities and relate to them.”
Why does this matter?
“If you are trying to integrate robots into shared spaces, you need to understand that people will treat them differently than other devices, and once you do understand that it is awesome because it is something you can try to harness,” Darling said. “Robots that have more anthropomorphic attributes seem to be more accepted when they are integrated into the workplace.”