This project brings together researchers from several disciplines with community partners to develop a range of novel sensing, monitoring, and messaging technologies to create a Virtual Safety Net (VSN) system for community members in an underserved urban community. In this effort, researchers will collaborate with members of the Black Ministerial Alliance and Health Ministry leaders of member churches, along with volunteers who provide health promotion outreach, the church leadership and a community liaison affiliated with a hospital to improve the overall health of this predominately African American community. The VSN is a research infrastructure for development of crowdsourced health interventions, in which community members author or culturally tailor intervention rules and messages. The VSN will empower the community to collectively solve health-relevant problems it identifies as important and provide preventive and health interventions that leverage support and messaging delivered individually via smartphone apps or during group meetings. The VSN will also be used to help the community identify and address social determinants of health, including food insecurity, homelessness, and health literacy. This research will have immediate impact for the 20,000 members of the 30 churches in the Black Ministerial Alliance of Boston. The technologies implemented can be rapidly disseminated to other church communities, businesses and social organizations in the US.

This SCC project aims to develop a research infrastructure that will be used to co-design and evaluate technologies in collaboration with community partners. The project will leverage the deep social support networks already existing in the Boston area. Key technologies developed in this effort include crowdsourced health interventions, in which community members author or culturally tailor intervention rules and messages. This will include developing methods to identify the range of modifications that laypersons can meaningfully make to expert-authored intervention dialogue and designing user interfaces to enable these modifications. It also includes advances in artificial intelligence-based indexing of community-authored narrative text for just-in-time messaging to motivate change in longitudinal, multi-behavior interventions. Advances in the automated generation of persuasive arguments for behavior change will also be developed. Dialogue-based interfaces to generate explainable artificial intelligence-learned models will be made to enable community members to understand and potentially modify how these models work.