Founder and chief scientist Cynthia Breazeal has been a pioneer in the field of social robotics.
Cynthia Breazeal is a pioneer in the field of social robotics—machines capable of recognizing and simulating human emotions. She developed her first robot, called Kismet, in 1999 as part of her doctoral thesis at MIT. Kismet is now on display at the MIT Museum with other early AI artifacts, and Breazeal has co-founded a company, Jibo, to commercialize a much-advanced social robot.
Breazeal envisions a future in which robots are adopted to improve people’s quality of life. “I can see social robots contributing positively in so many different ways,” she says, such as by providing personalized education to children, supporting the elderly in their own homes, and encouraging individuals with their health and wellness goals. “But before we can get there,” she says, “we need to introduce social robots in a natural, humanistic way while also building trust.”
Jibo unveiled what it calls the first “family robot” (also called Jibo)—designed to be an emotionally engaging, fun, and helpful companion in the home—on crowdfunding site Indiegogo in 2014. The project initially raised US$3.6 million from individual backers. Since then, Jibo investors, including venture capital firms Charles River Ventures and Flybridge Capital Partners, have provided another $60 million.
Putting the Emotion in Automation
Unlike other AI-powered automatons, social robots are built for engagement. “They prioritize the needs of the people around them and can forge bonds through a highly interactive presence,” explains Breazeal.
The latest Jibo prototype does more than respond to basic commands or questions. Using voice and facial recognition, Jibo can get to know the people with whom it lives, learning their likes and dislikes and becoming more helpful and accurate over time. The robot can recognize different people by voice and face and interact based on what it learns about each person.
And like its human companions, Jibo will seek out interaction. Rather than waiting for input, it will ask questions, make suggestions, or even crack a joke on its own. In the kitchen in the morning, for example, Jibo might greet a parent with, “Good morning, John. I hope you slept well,” while offering the children fun facts based on their interests.
“Today, technology treats us like technology,” says Breazeal. Robot companions must be more than transactional voice boxes; they must become relational and situational—capable of processing and responding with emotions that benefit the circumstances and tenor of different interactions.
Taking a Cue from Disney
While Jibo may behave in a more human-like way than other robots, it looks more like an animated, modern lamp than a humanoid. That’s a deliberate choice. Breazeal is keen to avoid what roboticists term “the uncanny valley,” the point at which robots appear so human that people are unsettled by them. “People enjoy robots that are anthropomorphic and emotionally engaging—but this does not mean the more human-like the better.”
“We have a natural fascination with being able to interact with and relate to the ‘not quite human other.’”
Breazeal and her team look to classical animation—think Disney’s anthropomorphic animals—to create robots that elicit positive human identification but don’t cross over the line to creepiness. “Disney is a great resource for understanding how to create compelling robot characters,” she says.
Thus, Jibo leverages key human social cues and features, like a cocked head, but the team reinterpreted them for Jibo’s physical form. “We have a natural fascination with being able to interact with and relate to the ‘not quite human other,’” Breazeal says. “It captures our imagination and connects to our humanity in a different way.”
It Takes an Ecosystem
Today, the biggest challenge for Breazeal and her team is delivering a product that meets her vision for a “valued, adored, and contributing member of the family” rather than just another piece of home hardware. The company is fine-tuning Jibo’s ability to interact with people in a natural, interpersonal way.
To further expand the robot’s capabilities, the company will soon release the Jibo Software Developer’s Kit so external developers can create applications that take advantage of Jibo’s body movement, screen animations, and voice. “We know that the potential for Jibo is profound,” says Breazeal. “Inviting third-party developers to tap into our platform to develop new content and skills for Jibo will provide our customers with even more value over time.”
Having scrapped earlier launch dates, Jibo now says its first robots will be available in late 2017. “Building the first robot for the home is a complex task, and we’ve made sure to lay the groundwork and build the proper foundation first,” Breazeal says. “This is an uncharted path that takes time and patience.”
Getting Out of the Lab
When Breazeal talks about social robots, she sounds less like an entrepreneur and more like an evangelist: “We have a responsibility not just to launch a product but to advance the social robotics category,” she says. She hopes the technology can “help us to become the people and society we aspire to be.”
After more than two decades at MIT, Breazeal felt compelled to introduce social robots to the mass market because she believes in their potential to transform early education, healthcare, and elder living. “But before we can get to a place where robots can step in and provide advanced support, we need to introduce social robots into our everyday lives,” she says.
Moving from the lab to an early stage company is a big change for Breazeal. “Research is about making scientific advancements and enabling opportunities that are beyond the reach of a commercial enterprise,” says Breazeal. Once you can sell the technology, the problem you’re solving changes. “We’re making social robots more accessible while also building a profitable business.”
Breazeal isn’t running the business day to day—Jibo recruited former Nuance executive Steve Chambers as CEO and Breazeal still works at the MIT Media Lab. But even as Jibo’s chief scientist, where her role is to monitor the development of the robot, assess its performance, and incorporate technological innovation, she must operate within time and resource constraints that are more stringent than in academia.
The challenges of making a saleable product at a profit are worth it, she says, because it will open the door to wide-scale adoption and development of social robotics to solve bigger problems. D!