Composable Hybrid Approach to Recognizing bots and their Influence in Social Media Analysis (CHARISMA)

Social media is transforming the way people around the world share messages, receive information, and come together to act. Unlike most traditional media, social media’s technical foundation enables messages to spread quickly and widely. This technology can also be used to disguise the true source of messages and produce automated messages that may have influential and manipulative effects on populations and communities. This vulnerability has enabled state, criminal, and terrorist actors to take advantage of social media to wage informational warfare. Forms of attack include spreading messages to influence audiences toward particular opinions or actions and flooding an information space with deceptive messages to prevent accurate information from being disseminated by legitimate sources. To detect and mitigate these threats, social media ethnographers and analysts require more powerful tools to filter, identify, and classify coordinated entities and their bots within large social media streams. These tools must detect threats, as well as distinguish between benign and dangerous behavior, while helping assess the vulnerability of human networks. To provide this capability, we propose to enhance and refine CHARISMA, a Composable Hybrid Approach to Recognizing Bots and their Influence in Social Media Analysis, which was developed under an ONR-funded Phase I effort.