Dr. Machiel Van der Loos, the Associate Director of the Collaborative Advanced Robotics and Intelligent Systems Lab (UBC), forwarded me an article on Paro from the Wall Street Journal a couple of days ago. Paro is a furry baby seal robot with the potential to impact people emotionally: it has two huge eyes that can stare at you adoringly and it can move its body parts like a real pup. Emotional bonds with artifacts, such as objects, etc, have been discussed for hundreds and thousands of years in literature. Greek mythology alone provides a few relevant stories on this very topic, not to mention more modern tales, such as The Adventures of Pinocchio. More and more of these stories are becoming a reality via robotics technologies, but not all have happy endings. The question for discussion today is, ‘Would our emotional bonds with Paro lead to a happy ending, or an ethical ending?’
I had the chance to interact with a Paro robot when I was at the 2009 International Conference for Robotics and Automation (ICRA) to attend the biannual Roboethics Workshop. I had just graduated with my BASc from the University of Waterloo, and was ‘pampering myself’ for the years of hard work as an undergrad by going to the conference and traveling around Japan – nerdy, I know. Eager to do anything and everything possible to learn about the current state of research in robotics, I restlessly hopped from one seminar room and exhibit booth to another throughout the week-long conference. One of the booths was exhibiting two Paro robots, next to another booth housing the Keepon (BeatBot) social robot, and I was instantly drawn to them. I waited in line during session breaks to interact with both Keepon and Paro.
While I inquisitively asked technical questions at other exhibition booths, my questions about Paro were quite different. Now that I think of it, I don’t remember asking too many questions at all, perhaps because I was more interested in interacting with the robot rather than the person introducing it to me.
That is precisely the ethical issue people are starting to talk about.
According to the Wall Street Journal article by Anne Tergesen, Paro is already an official medical device in the US, categorized into Class 2 along with powered wheelchairs. Paro robots are mainly being sold to nursing-homes throughout Japan, Denmark, and the US. Paro, which apparently cost $15 million to develop, is known to have positive impact on the elderly. Ann Tergesen gives examples of its soothing effect on people with dementia. Apart from that, Paro is known to encourage a more social atmosphere in elder care facilities by naturally promoting more interaction between the residents. In one of the video clips provided by the manufacturer, a group of typically introverted seniors gradually began interacting with each other more when Paro was around. One of the residents even tried to feed her snack to it.
Yes, the lady in the video tried to feed the robot.
You might consider this totally unusual behavior. But it goes to show the extent to which people, especially emotionally and socially vulnerable individuals, such as the population of an elder care facility, anthropomorphize (or, in this case, zoomorphize) social robots.
At ICRA 2009, Dr. Mattias Scheutz, an Associate Professor at the Indiana University and the Director of the Human Robot Interaction Lab, presented his work on “The Inherent Dangers of Unidirectional Emotional Bonds between Humans and Social Robots”. Similar to Dr. Sherry Turkle in her warning against regarding Paro as a companion, Scheutz is not entirely jumping with joy about the emotional bonds people create with these social robots. He believes that an investigation of potential harm and other effects of the machines is essential due to their power to prompt humans to ascribe agency. He also worries about the potential of these robots to exploit human emotions:
“For these personal interactions might prompt or cause humans to form, for the time and state-of-the-art being, unidirectional emotional attachment relationships with robots that are not only not appropriately reciprocated, but might allow robots to take advantage of people’s emotional propensities and reactions.”
Scheutz’s concern is certainly not singular. Turkle asks “Why are we so willing to provide our parents, then ourselves, with faux relationships?” In the same vein, Dr. Christoph Bartneck suggests (through his thought-provoking poster) “Create Children, Not Robots”, meaning that we should focus on increasing birthrate to produce larger younger population rather than creating robots to meet the shortage of healthcare workers caring for the elderly. Arguing in favour of social robots, Levy counters by hypothesizing that love and sex with robots is inevitable and will be regarded as natural – or at the very least widely accepted – in the future.
For the time being, some organizations that have started to use social robots seem to remain cautious. While encouraging the use of Paro in nursing homes, the Danish Technological Institute is taking actions towards investigating its effects on the elderly:
“DTI requires caregivers to attend Paro seminars, where they discuss such issues as whether it’s OK to leave an elderly person alone with a Paro, and whether patients must be told it’s a robot. Don’t allow someone to “escape into a strange seal robot’s universe,” cautions Lone Gaedt, senior consultant at DTI.”
As someone working closely with the effects of social behaviors of robots, my research presentations usually start off with a motivational slide suggesting the necessity of building robots that can naturally interact with humans and follow our social norms. I strongly believe that developing social behaviors in robots is essential if we want them to be truly useful to human society. Using them as aides in caring for the elderly is just one example of how robots can fill certain social roles. We certainly need a means of taking care of the aging population in the near future, and we clearly aren’t showing skyrocketing birthrates today.
At the same time, I also believe that replacing the traditional senior-nurse (human) interaction with senior-robot interaction could undermine the beauty of how a younger population obtains wisdom from its elders. Robots, using their programmed artificial intelligence, can expand their knowledgebase and customize their behaviours to meet user needs. But humans not only obtain knowledge for the sole purpose of better serving the other person; rather, they infer wisdom and learn life lessons that can then be applied or passed on to others outside the context of a nursing home.
As I wrote that last paragraph, I remembered my grandmother who suffered from a stroke and constantly needed someone (that someone occasionally being me) to look after her. The memories of my grandmother bring with them a sense of appreciation for the times I spent with her while she was alive, mixed with a hopeful thought that maybe the person she helped me to become today can somehow have a positive impact on the world.
These mixed feelings aside, I have to admit that most of my time at the exhibit booth was spent saying, ‘Awww!’, petting Paro (while my obsessive compulsive self wondered how many people must’ve touched the robot and made a mental note to wash my hands afterwards…I found out later from The Wall Street Journal article that the fur is actually made of antibacterial material) and giving it adoring looks. So I can easily understand Ms. Simmeth from the article, an introverted resident of a care-facility in Pittsburg:
“I love animals,” explains Ms. Simmeth. She whispered to the robot in her lap: “I know you’re not real, but somehow, I don’t know, I love you.”
Video clips on Paro (mind you, interacting with the robot in person and watching a video clip of it is an entirely different experience):