MarvelousAI is an early stage startup founded by tech industry veterans. We are building an augmented analytics platform to provide actionable insights regarding online narratives. Our “Cyborg” intelligence methodology detects narratives and fine-grained emotional content, combining the power of human-in-the-loop learning with the latest in natural language processing, computational linguistics, and machine learning.
The first use case centers on US electoral politics, specifically the 2020 race for President of the United States. Target customers are political campaigns, journalists, academics, and concerned citizens. We are particularly focused on how messaging spreads among and between diverse online media, both negatively (propaganda) and positively (truth-based messages); which narratives persist over time; and the impact they have on overall discourse about a topic. Our StoryArc platform has been used to discover and quantify political bias, asymmetries in coverage of male vs. female candidates, and active misinformation campaigns.
The second use case focuses more broadly on Misinformation Security and Responses. In collaboration with the Credibility Coalition, we are developing a standard framework to define the equivalent of the MITRE ATT&CK Matrix for Misinfosec. As a practical application of the framework, we address a broad spectrum of areas such as US immigration; health policy and impact; and climate change.
Our methods can be applied to detect narratives and fine grained sentiment related to any topic area, language, or source.