Hailed by some as a genius, Ishiguro is an engineer who has been working at the frontline of robotics for over three decades. Crafting what most see as the worlds most human-like androids for critical mass usage. These robots form part of a larger technological trend that is aimed at helping not only the human species with arbitrary tasks like cleaning and cooking but more interpersonal duties like partnering with lonely people or assisting the cognitively impaired (Alzheimers) with high-quality conversation. For some, these creations elicit revulsion, for others a sign that the revolution is about to begin.
For Hiroshi, each and every day he spends with these technical creatures, he learns just that little bit more about what makes us human. Some of his creepy creations have been modelled on himself (see below), others like the prototype Erica are a hint of what an artificially sentient human might look like. At some point, Ishiguro will need to get his creations past the uncanny valley threshold (life-like revulsion) for us to live comfortably with them, but for now, these robots are part of a larger intellectual dream which Hiroshi is more than happy to investigate with me.
Hiroshi Ishiguro / photo: Makoto Ishida
What would you say is the most important thing that we don’t understand yet about humans?
Emotions and consciousness, they are both critical topics to look at over the next decade; this is an area that we need to study a lot more. People consider emotions and consciousness to be based around intention and desires, but we don’t know the exact definition. If we look at neuroscience, they don’t understand the precise mechanisms for emotions and consciousness. If we create a human-like robot, and we can feel consciousness and emotions through interactions, I believe we will have hints in order to understand this.
This is a unique paradox in terms of the human condition that we don’t understand ourselves, but we can perceive others quite well. Is this what you mean?
Yes, but to understand myself, I need to interact with other people. I cannot observe myself; it is only through my interaction with other people that I have a chance to understand myself. Human motivation has always been about observing the external world, not the internal world.
Hiroshi, you’ve said that you believe in 1000 years all humans will be robots. Is that true?
Possibly. Look at disabled people, they have prosthetic arms and legs, but we think of them as humans. We never define humans based on their bodies. That is according to our understanding. In the future, I believe we’re going to replace our organs with artificial organs, and then we will be able to survive for longer, even if severe situations arise on our planet or on the sun. If we can replace our bodies with machines, then we will be able to survive in space. But 1000 years is too short; I want to say that it will be more like 100,000 years.
"If we can replace our bodies with machines, then we will be able to survive in space."
Hiroshi Ishiguro on our future robotic life
We’re talking about the definition of a human here. They’re creating artificial wombs. People are biohacking. People have exoskeletons; we have prosthetic limbs; people have pacemakers. All of these things are creating a hybrid form of a human a type of artificial life?
Shortly, we may have implants, and we may connect our brain with computers. But still, I believe we can have a human-like consciousness where we use computers as a part of our brain. We don’t need to care about the materials, and we don’t need to care about new organs if we have a computer where we can simulate human-like consciousness.
In the work that you do, how far away do you think we are from having a mirror-like human?
We need to do so many things, and our project is a starting point for the development of human-like robots. It’s challenging to answer that question. But I would recommend you watch our movie with Erica. Erica is the most advanced android, and in limited situations, Erica can talk with visitors for maybe five or ten minutes. Upon first contact, we never experience a deep conversation. We usually talk about limited stories. Erica can talk to some visitors upon first contact, but we need to stretch the limit and purpose of the robot in every way.
You have talked about sex robots as well as robots helping with cleaning and robots helping the elderly. There are a lot of experiments that are being rolled out around the world within this area. Do you think humanity is ready to live side-by-side with these assistant robots?
The technology is possible, but of course, we need to choose the situation and the purpose carefully. We can design the robots, but we cannot have a general human-like robot.
Are you saying that this is something we shouldn’t have?
It depends on the morals of our society. We need to consider how to use human-like robots carefully. What is a useful purpose, and what is not? Do we need to have a higher level of morals when using powerful technologies? Always.
What situations or applications do you think would not be suitable for robots right now? Where would you not put them?
Of course, we should not limit its purpose. Still, we should also be aware that robots can be used for dangerous purposes such as cheating people, however for people who want conversational robots such as using it for people with dementia and autism; they need a robot and these kinds of things are already happening with computers.
"We plan to construct human-like robots, and if we can see some humanity there, then we will be able to understand humanity better."
We seem to have an anthropomorphic obsession. This idea that everything we create should be carved out of our concept of what a human is. Do you think that’s the right way to approach our relationship with technology?
I think it is a matter of our human brain function. Our brain has many features to recognise humans; therefore, an anthropomorphised robot is better for social interactions. It’s easy to interact with people, but again it depends on the situation and the purpose. Sometimes it’s better to have a human-like robot; sometimes it’s not. With autistic kids, they don’t like to have a very human-like robot; they want a more cute and doll-like robot. So we need to design the robots based on the purpose and the users.
Are you saying that certain demographics in society would not want robots to look human but look like something else?
It’s better to be anthropomorphised, but there are many levels of anthropomorphism from very human-like robots to just very simple robots. But a robot is a robot.
Who do you work with to understand humans? We’re complex creatures, and you have to transfer this knowledge into the robots. How do you approach this?
We don’t know what the most effective and efficient way to understand what a human is. We’re working with neuroscientists and social scientists, and they’re using all of our work. This approach is a combination of traditional research areas for understanding humans and robotics.
What do these conversations look like?
It is not so difficult. For example, we’re using our android for neural feedback. It’s a way of controlling the brain signal by giving input from the computer screen. If we use the android bodies for the neural feedback, we can have more powerful control for the human brains. So we don’t have any problems with discussions or conversations. We’ve already spent around twenty years working together, and I have many excellent collaborators. They understand the robots well.
I was at a science festival yesterday; I saw a discussion between a Buddhist monk and a neuroscientist, they were talking about what makes a human distinct, what is consciousness? How do we define the self? The monk was talking about the investigation of the self, whereas the neuroscientist was talking about defining the mechanics of the brain. Do you sympathise with this difference?
The monk tried to understand the brain by using his mind but the neuroscientist attempted to understand the brain mechanism right? Unfortunately, the brain is so complicated, nobody knows the mechanisms of the brain yet. And my standing point is maybe in between.
What do you mean?
By creating the human-like robot, it can reflect humanity henceforth we can understand humans better.
So your work is all based around reflecting humanity?
This is a possible approach, I believe. I call this approach the constructive approach. We plan to construct human-like robots, and if we can see some humanity there, then we will be able to understand humanity better. Neuroscientists try to understand our functions based on the brain mechanism and the monk attempts to understand this by using their brain. I’m trying to understand by creating a human-like robot.
Would you ever work with a monk alongside your work? Maybe they could understand a bit more about the human psyche or the experience.
Yes, possibly, I’m attending that kind of conference, and I’m giving a lecture, so we’re running a lot, but monks are not technical. We can get some hints, but I think we cannot clearly understand what a human is based on their approach.
You would like to be more empirical?
Yes. I can get some inspiration because they’re talking about humans from a different way, but that’s all in some sense. I can work with neuroscientists better.
Why are you working with a neuroscientist if we don’t understand the brain at all?
It depends. They’re taking a bottom-up approach. Our approach is top-down. So both methods should be integrated.
Perhaps Japan is a culture that promotes and motivates you a bit more to be a roboticist. Japan has the oldest population in the world, one of the highest suicide rates, 30 per cent of Japanese men in their 20s and 30s have never dated. Do you think that some of this culture inspires you to create the work that you do?
I don’t know, but as you know, in Japan, we are very open to robotic culture, we never distinguish between the human and the robots. That is our tradition. Japanese people have a big family, and in the family, we embrace everything equally.
For Japanese culture, is there a more profound need for robots than in other countries, considering where the economy and the society is?
In Japan, we have a very different situation. Our population is going down a third of its current size within 50 years. Therefore we need to have more robots.
"Our population is going down to a third of its current size within 50 years. Therefore we need to have more robots."
What’s the next big thing that you’re working on?
I want to change this world by creating interactive robots. As you know some companies, they tried this, but still, they didn’t get big success. I need to be among the players. Now people are using smartphones, but I think we may accept interactive autonomous information media like a conversational robot.
Is there a robot war going on right now? Are you competing with China or the US on this?
I don’t know what kind of thing is happening in the US and China, but if we get a big success in Japan, then they will try to make a copy.
But generally, we don’t think about our needs before developing new technologies. Today we have a lot of techniques, but the technology may change, for example, iPhones. Nobody said we want to have iPhones. But when Steve Jobs developed the iPhone, everybody accepted that technology. These are the typical steps: style, needs, technology development and then social acceptance; they all happen simultaneously.
What do you think android’s think and do they dream?
Well, not yet. But that is the next big step for us. One of the essential features of a human brain is imagination. We’re using imagination for controlling our bodies, for taking action. The question is how we can implement that kind of function for the androids. Of course, this is the next big challenge.
We need to give robots consciousness, right?
Consciousness and imaginations.