HCC: Small: Understanding How Teens Use Generative AI Agents

Teenagers are increasingly using tools that depend on generative artificial intelligence (generative AI). Social media companies are rapidly integrating generative AI into their platforms to create agents that they describe as assistants, companions, and coaches. However, research on how the use of generative AI agents affects teen well-being lags behind teens' exposure to these systems. In this project, the research team will study how teens are interacting with generative AI agents today and how this use impacts teens' well-being. The team will also explore how technology designers can use this knowledge to create more useful, safer generative AI agents for teens. The contributions of this work will include new scientific understanding of how teens are using generative AI and new guidance for maximizing the benefits of these systems for young people. The research is organized around three main activities. In the first activity, the team will recruit teenagers who use generative AI who are willing to share selected parts of their interactions with these systems. Through analyzing the chat histories and using them to guide personalized interviews with teenagers, the team will develop a nuanced understanding of what kinds of conversations teens have with generative AI systems that impact their well-being. In the second activity, the team will ask therapists who work with teens to themselves interact with generative AI systems and review findings from the first study. This will allow the research team to develop an expert perspective on the benefits and risks of current generative AI systems as well as design principles for systems that are more likely to improve teens' well-being. In the final activity, the team will conduct co-design activities with teens to create generative AI agents based on the new knowledge and principles developed. These agents will be assessed through a comparative field study where teens interact with these agents versus commonly available ones and report on their interactions and well-being during the study. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.