Primary Investigator:
Gene Spafford

Project Members
Nicholas Harrell

The rapid proliferation of online discourse, particularly within social networks, has increased the spread of information at unprecedented rates. Although viral content often captures public attention, the underlying mechanisms driving its dissemination on social media remain illusive. This study highlights how value-laden features can be used to detect the potential of certain discourse on social media to become viral. Using a combination of Natural Language Processing (NLP) techniques and social network analysis, this research identifies patterns of user engagement that predict the likelihood of content achieving viral status. This study shows statistically significant differences in the value profile between the high and low engagement profiles of many topics on different alternative social media platforms. The results are further validated through predictive machine learning models. These results with discussion contribute to ethical concerns and implications related to the implementation of contemporary and future AI technologies that are being used for purposes of influencing public discourse on the Web.