In Hollywood films, robots are usually portrayed as extremely intelligent and human-like – but rarely benign. Indeed, at some point during the movie, you can bet they’ll blow a fuse and try to take over the world. So as AI and android technologies develop in real life, no wonder the concern is that robots could become too clever for their own good.
Yet, in this installment of The Debate, two distinguished professors of robotics – Gordon Cheng from the Technical University of Munich, Germany, and Hiroshi Ishiguro from Intelligent Robotics Laboratory at Osaka University, Japan – reveal that, while huge strides are being made in this field (both have created robots with human form), we are some way off from developing an emotionally intelligent robot.
Even so, we’ll increasingly see robots (humanoid and otherwise) in our homes and workplaces, and must learn to accept them, and even work in partnership with them. Professor Cheng and Professor Ishiguro highlight what this human-robot interaction could look like, and where we might encounter them in the years ahead.
Professor Hiroshi Ishiguro
A robot is an extreme example of technology. Those of us working in robotics always want to push at the technological boundaries because it’s a way to expand the realm of human possibility. Robots and AI are mirrors that reflect our humanity and can help us explore big questions, such as: “What does it mean to be human?” and “Why we are here?” This is why I developed Erica, a fully autonomous android.
In the past, Hollywood has been responsible for much of the negativity surrounding robots because, in films like “The Terminator,” they’re always destroying the world. I think this is a big cultural difference between America and Japan. In my country, the robot has been seen as a friend from the very beginning. Having said that, Hollywood is now crazy for the technology and, in a couple of years, there are plans for Erica to star in her own $70 million sci-fi movie.
Still, at the moment, Erica’s abilities are quite limited. She sits in the lobby of our research institute and talks to our visitors and, because she is friendly, people accept her as “human-like.” That says a lot about how accepting we are as a society – although Erica is very far from being human.
As a child grows, they develop more complex conversational abilities, and Erica will do the same. As her capabilities improve, she’ll have an even better relationship with humans. Of course, if we want to create a perfect robot copy of a human being, we’ll need a leap forward in technology. Yet many interesting things are happening right now, such as computer scientists improving computational powers, and materials scientists improving robot skin.
One of my challenges will be to give Erica a more natural, human-like intelligence – although the big difference between artificial intelligence and human intelligence is that AI is based on big data, whereas human intelligence is derived from our interactions with the world around us.
I believe robots will enrich society. For example, we did an experiment to find out whom people would prefer: an android shopkeeper or a human shopkeeper? Actually, android shopkeepers were very popular, especially on the men’s floors of department stores. Japanese men dislike to talk to human shopkeepers when buying clothes. We are shy. It is not the samurai spirit. Samurai spirit is all about not caring how you look or dress. If a human shopkeeper says, “That looks good!”, we never trust it. But when androids say the same thing, we accept it, because androids never tell a lie.
Erica, a humanoid robot, converses with a team member from ATR.
Professor Gordon Cheng
I am biased, of course, but I believe AI and robots can help us become more productive and live better lives. Robots can certainly fill productivity gaps in areas such as hospitals, logistics and farming.
For instance, I was talking to a farmer in my home country of Australia recently. Because of lockdown, vocational workers were unavailable to help pick fruit, so it was going rotten before it was off the tree. That’s a massive waste, so the longterm solution is to get robots to pick the fruit instead. It’s also more efficient. As an example, an olive farmer told me that robots can detect which olives have parasites and which don’t, something that’s difficult for humans to do. I also spend time talking to doctors about the way healthcare is delivered. We can modernize a lot of treatments with AI and robots to give patients a better experience.
Meanwhile, the logistics industry is under pressure because of a massive staff turnover every year – so this is a sector where robotics will increasingly come into play. Staff shortages are not just an issue in the western world, either. Chinese factories also have to deal with this problem, and have brought in automation to stop supply chain disruption.
There is, of course, the old argument about robots taking people’s jobs. But the point is that people don’t want to do dirty, low-paid or unsafe jobs – such as cleaning the outside of high-rise buildings – and they shouldn’t have to!
There are two ways that robot-human interaction can develop. One is that robots will simply do what we tell them, much like Alexa or Siri do now. The other is that they will learn our habits, and become more like companions. Personally, I think robots should have some form of emotional intelligence (EQ), otherwise our interactions with them won’t be meaningful and the robots will just be reciting a script. But creating EQ in robots? That’s a major challenge.
People might find robot-human interaction difficult to come to terms with at first, but remember: 100 years ago, people adapted to seeing motor vehicles – rather than horses – in the streets. A year ago, not many of us were using Zoom, but now most of us use it, every day. Yet I’m not for forcing change on people suddenly. I’m for slow adaptation. Nowadays, I think engineers can be too forceful. They make certain assumptions and expect the user to immediately conform to it. I don’t think that’s the best way to introduce any kind of meaningful interaction, particularly with robotics.
Ultimately, we have to ask ourselves: How do we get human-robot interaction working in a bidirectional and meaningful way? And how can we use robots to best serve human intention? Those will be our biggest challenges.
Published: June 2021
Images: Getty images; Hiroshi Ishiguro Laboratories; private
Video: ATR Japan; ERICA: ERATO ISHIGURO Symbiotic Human-Robot Interaction Project