Episode Summary: In this episode, Stephan draws on his robotics background to articulate what it takes to give a robot a “personality”, explaining the differences between responses and propensities along the way. Androids are already making news in the entertainment and retail industries, but we delve into why the health sector is one of the next big industries, and how culture might influence social acceptance across country lines.
Guest: Stephan Vladimir Bugaj
Expertise: Robotics Design and Filmmaking
Recognition in Brief: Stephan Vladimir Bugaj is a modern visionary with extensive experience in screenwriting, technical artistry and directing in animation and games. He is the Creative Director at Hanson Robotics, where he specializes in robot personality and functional design. He is also a writer-director for WakingUp media and Visioneer studios, two screenwriting and production companies, and part of the story “brain trust” for Limitless VR. Stephan worked for over 10 years as a screenwriter and technical director with Pixar Animation Studios, and before that was a multimedia researcher at Bell Labs and artificial intelligence developer at Intelligenesis/Webmind.
Current Affiliations: Creative Director at Hanson Robotics
Giving Robots a Personality
How do you give a humanoid robot a “personality?” This is not a rhetorical question, but it is one that requires some deconstruction of the concept of a personality before we can arrive at an answer. When I sat down with Stephan in a recent interview, he explained how Hanson Robotics is conquering this task.
“Our approach at Hanson Robotics is to look at the semantic, cognitive, and functional aspects of personality, and figure out how to encode those in robotic systems and software,” Stephan explains. Hanson is specifically interested in how personality is expressed and responded to at a low and high level, but from a functional rather than a more underlying, architectural perspective.
Stephan describes personality as a characteristic of an intelligent entity, encompassing (albeit abstractly) expression of self and response to various stimuli. “At a less broad level, if someone is mean to the robot, is it mean back, or does it turn the other cheek.”
When the Hanson Robotics’ team programs responses, they are always cognizant of the fact that the robot needs to be very expressive and lifelike. The ultimate goal is to have people interact and feel comfortable with the emotional responses that they are receiving from a robot. When I throw out the term ‘propensities’ in the same sentence as ‘responses’, Stephan separates the two. “Propensities are much more interesting and are definitely more of the direction we’re going in, in the immediate and longer-term”, he clarifies.
“An emotional model for a robot would be more along the lines of weighted sets of possible response spaces that the robot can go into based on a stimulus, and choose a means of expression within that emotional space based on a bunch of factors.” A robot with propensities would have a set of different considerations to think about i.e. “What do I think of the person? How did it act in the last minute? How am I feeling today?”, which is more in line with the barrage of thoughts and feeling that filter through a complex human brain before every intentional thought and action.
A variety of contexts around immediate stimulus would then set an emotional frame that allows a robot to have a more complicated response to each of the stimuli, in which it (hypothetically) uses recent memory to build a longer-term emotional model.
“You can think of it as layers, you can think of it as interconnected networks of weighted responses…as collections of neurons, there’s a lot of different ways of looking at it, but it basically comes down to stages of filtering and considering stimuli, starting with the input filter at the perceptual level.” Humans filter out this extra input all the time, as they focus attention on one task or object and filter out most of the other information in the room.
A robot, like a human, could have more than one response to a stimulus. They may have an initial reaction or reflex, but then a few seconds later take a more “considered response” based on spreading of information through a neural network. Stephan gives the hypothetical situation of a friend coming into the room and taking a swing at another friend. At first, the friend on the receiving end might react by bolting upright and assuming a fighting stance, but a few seconds later he or she may realize that the other person might just be playing around. Hence, the stages of an emotional reaction and the making of a personality.
Basically, perception, context, and analysis are the tripod foundation of a considered response on the part of a complex being, which potentially includes robots. Today, the robotics team at Hanson is modeling these emotional robotic models after complex cognition, but at a more rudimentary framework for the time being.
Will We Mingle with Robots in the Near Future?
The progress that we make in 10 years with robotics (or any field), is always speculative, but it is always worth considering where we are now and the directions in which those fields may be moving.
At Hanson, the team is moving toward affective robots, those that have emotional relationships with humans (and we’re not talking about sex bots). Stephan notes that the entertainment applications are obvious, being steeped in this line of work himself. Entertainment speaks not just to film-making but also to robots serving as greeters at places like theme parks and trade shows – anywhere that having a relationship with a human would be of value. “Androids are designed to relate to people and develop emotional rapport,” explains Stephan.
He believes that the relationship component naturally leads into educational and healthcare applications, and other areas such as hotels and high-end retail. He mentions a survey conducted in Japan, in which greater than 50 percent of elderly responders stated that they would prefer to be taken care of by a robot than a young person. This is a fascinating phenomenon, and one that is presumedly steeped in cultural context.
It could be that Japan, who is ahead of the curve in robotics, simply has the capability to make something that could actually perform this task in the relative near future. Either way, Stephan believes we will see these types of service-oriented and emotionally-supportive robots in Asia and more progressive, perhaps more urban parts of the U.S. in the next 10 years.