CAREER: Redesigning Engagement Infrastructure On Social Media For Well-Being

Social media is often blamed for making mental well-being worse for adults. We have a general understanding of how platform design and use patterns can cause these problems. However, there is still much to learn about the mechanisms by which social media use affects well-being and what people and platforms can do to make things better. This project addresses these challenges by developing better models of how people use and respond to social media content, along with new AI-driven technologies that improve platform design and the algorithms that choose what, when, and how to show content to users. These new models and technologies will be guided by a wide range of studies of social media use that will enable early detection of possible harm to well-being. These detection models will be integrated into tools that help people think about the benefits and risks of using social media, as well as design. Finally, this project will create new designs for social media that reduce harms to well-being. To increase real-world impact, the research team will also engage in a multi-part public engagement effort to share the scientific results of the work with industry, policymakers, and the general public. This project develops new scientific knowledge, methods, and design interventions to understand and create more beneficial social media "engagement infrastructure" -- the interface and algorithmic features of social media platforms that help shape people's behavior and well-being. The research team will first conduct several studies to empirically characterize and precisely measure the engagement infrastructure that induces harmful use, developing a validated typology of the most problematic elements and mechanisms of action. These insights will be used to create and evaluate multimodal AI models for inferring when harmful use is occurring. Finally, the research team will design new interventions that interrupt and displace harmful use, either through platform-level changes or, for individual use, through overlay tools that change the engagement infrastructure. This approach will enable the decomposition of the precise mechanisms that mediate negative well-being outcomes, advancing human-computer interaction, social computing, and applied AI research. The project's outreach and education plan includes bi-directional interdisciplinary workshops at a number of venues to bridge research gaps, academic-industry collaborations for networking and knowledge transfer, public engagement through Minnesota State Fair exhibits and science communication partnerships, and curriculum development to train software engineers on responsible and ethical design. 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. NSF Award ID: 2540768 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT,01002930DB NSF RESEARCH & RELATED ACTIVIT,01003031DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Stevie Chancellor | Institution: University of Minnesota-Twin Cities, MINNEAPOLIS, MN | Award Amount: $432,014 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2540768 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2540768.html