While much work has been done in recent years to understand the digital media practices of youth—including research focused on learning and motivation, interest, identity, social development, and creative expression—little has been done to understand how advances in data science and artificial intelligence might be leveraged to increase the quantity and quality of child-child and parent-child digital play in middle childhood. Studies suggest that the well-being of youth is determined, among other factors, by the quality of social experiences they encounter online. Critically, brain science research has shown that the quality of a person’s relationships and social interactions shape their development and health, both of the body and of the brain. Whether gaming with friends, hanging out on their phones, or playing together in a virtual world, youth are interacting with others in social settings. While worrying to some, online participation can be harnessed to help youth build knowledge and skills that are important to them, and that help them become fully engaged in the world around them. This is especially true during middle childhood, a period marked by a rapid convergence of significant cognitive, neurobiological, emotional, and social transformation. It is also a time when youth experience increased social interest and independence, and when parenting behaviors can both support and hinder these developmental processes.
Data Science for Digital Play is a multi-year initiative bringing together research scientists and developers in game design, artificial intelligence, machine learning, data and learning analytics, and social network analysis with experts in child development, community psychology, interaction design and children, learning, and play. The goal: develop novel automated approaches to the research and design of play-based online environments focused on increasing the quality and quantity of parent-child and child-child digital play in middle childhood. How might machine learning algorithms be used for studying child-child interactions in Minecraft or Lego Life, for example? How might natural language processing be used to better understand the kinds and qualities of social interactions children are having in Roblox? How might data analytics inform the design of online platforms that encourage more parent-child digital play?
Project Lead: Katie Salen