CHS: Small: Exploring and Predicting Unintended Consequences of Technology

University Of Washington, Seattle WA

Investigators

Abstract

This award investigates the unintended consequences of Internet technologies that people have experienced in the past and that may arise in the future. Leveraging the "wisdom of the crowd," the work will give a voice to people across the US, including underrepresented populations, by surfacing how technology products have affected themselves and their communities - both positively and negatively - beyond the intended purposes. The work will also provide tools and predictive models that help product teams anticipate the side effects that new technologies might have on specific user groups and their communities. The research includes (1) a large-scale study on the principal investigator's LabintheWild platform to elicit qualitative reflections on a set of existing technology inventions; (2) automating the analysis of this data using unsupervised structural topic modeling to determine how different demographic groups perceive current technology and to relate themes of unintended consequences to specific categories of technology; and (3) developing a volunteer-powered, data-driven tool that provides a lens into the views of broader society and supports companies in assessing and foreseeing potential unintended positive and negative consequences without needing to reveal proprietary ideas. The tool will automatically update its predictions based on newly collected data to remain useful as society moves on and new technology inventions appear. Educational components include a new course module on societal implications and ethics of technology, involving students in all parts of the research, and visualizations of unintended consequences that will be developed for scientific outreach at middle and high schools and for public access. 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.