Robots for Social and Emotional Learning
Haru as a Social Mediator between Cultures and Generations
Haru and other social robots have the capacity to act as a peer and mediator of activities that bring people closer together. Through multiple projects, we explore the ways Haru can act as a bridge across continents and generations, helping children build their their interpersonal social skills, their global citizenship skills and their awareness of and sensitivity to people who may be different from them in age, cultural background, and daily experiences.
In the first project, we utilize Haru as a mediator between classrooms in different regions of the US and countries around the world. Through remote communication, children can engage with Haru’s activities and share their everyday practices and experiences with another classroom, to inspire multicultural understanding and a feeling of belonging to a larger international community. While children connect with each other virtually, Haru supports their communication and sharing through its embodied presence in both environments. Haru enables both synchronous interaction, and asynchronous communication between children in geographically separated classrooms. We work with Honda Research Institute and other collaborators in Japan, Australia, Belgium, and Greece, connected through the Socially Interactive Robots Consortium (SIRC), to deploy this system internationally. In the US, we partner with diverse local rural and urban schools, such as Batchelor Middle School and Edgewood Junior High School, to develop and evaluate design concepts for educational robots that can support student learning and emotional growth. Our future plans are to expand our network of participating classrooms more broadly in the US, and in other countries around the world, including Bosnia, Turkey, and Morocco, among others.
In a second project titled Intergenerational Music, we explore the way Haru can mediate music activities for young children and people living with dementia. We bring Haru into an intergenerational daycare to incorporate these groups in the development of Haru’s activities, learn about the synchrony between Haru and its users, and tailor Haru for music related activities with a diverse audience. We co-design appropriate activities and ways to measure their effects with staff and residents at Jill’s House Memory Care. These activities are meant to serve the broader community goal of Jill’s House to provide older adults and children with opportunities to express themselves creatively and improve their social and cognitive skills through intergenerational interaction.


Robots in Family Homes
Haru4Kids (H4K) is an app-based robot designed to cohabitate with children in their home. Our work involves collaborating with other SIRC partners to develop appropriate content that can keep children engaged with the robot over longer periods of time, such as multiple weeks, months, and eventually years. The content is engaging for children, as well as educational, aimed at helping them develop their verbal and computational skills, as well as their knowledge of subjects that might be relevant to their cognitive and school level. Through this home-based socially interactive robot system, we also hope to expand the accessibility and inclusivity of educational content, which children can have access to in a space and time that is convenient for them. Pre- and Post-interviews with caregivers and children allowed us to gauge familial comfort with sharing different kinds of information with the platform, and to assess how their experiences living with the robot changed their comfort levels. Our work suggests co-habitation robots need to incorporate rich narrative-based activities and a wide variety of content to keep children’s attention over time. Child-centered, robotic design must incorporate children’s feedback, as parents may not be aware of children’s preferences. Our work will be published in the International Conference on Social Robotics in December 2022.

Robots that Support Children’s Rights
Under the framework identified in UNICEF’s policy guidance of children’s rights and AI, we explore children’s priorities and concerns in the context of social robots as technologies that can mediate between the digital and physical world. To do so, we conducted a series of four workshops within a variety of children’s communities which focused on introducing children to their rights with AI (session 1), fairness and non-discrimination (session 2), inclusion (session 3), and privacy (session 4) within the context of child-robot interactions. Children’s diverse perceptions provide us with child-centered recommendations for designers of child-robot interactions. We continue to iterate on these workshops to empower children’s voices.
These workshops also aim is to provide children with an opportunity to think critically about emerging AI and robotic technologies, and to creatively engage in imagining future technologies that would serve their needs and priorities. By considering the potential uses and consequences of these technologies for children in different contexts (e.g. rural versus urban, different countries), the workshops also foster children’s ability to expand their perspective on the world and their empathy with others. Finally, the workshops aim to help children develop their self-efficacy with technology as it affects them and their peers around the world.


Inclusive Sensing and Engagement
In this project, we present a multi-modal exploration of the stages of engagement, with a focus on developing modes of perception and engagement measures that are inclusive of diverse children (e.g. ethnically, neurologically). This exploration was made possible through a Robot Operating System (ROS) setup with visual, thermal, and audible sensors to collect data while children interacted during a passive and active activity with the robot, Haru. Currently, we are exploring preliminary results and using social behavior annotation tools to achieve a more holistic understanding of how a multimodal approach can increase our ability to approach engagement estimation in a non-invasive, contextualized, and embodied way. This system is envisioned as a foundational component of future social robot design, with the aim of improving the accessibility and inclusivity of robotic systems to be able to support interactions with a wide variety of children with diverse learning and interaction styles.
