TYBALT: Transcending cYber Barriers with Automated Language Tracking

TYBALT (Transcending cYber Barriers with Automated Language Tracking) primary goal is the automatic detection of adverse intentions against US forces via publicly available information (PAI). Our “Intent to Act Adversely Detector” (ITAAD) is based on empirically-driven cognitive models linked to features detectable in the structure (rather than content) of written language: Integrative Complexity (IC) - a model reflecting the ability to integrate contradicting perspectives - and Moral Disengagement (MD) - a theory of linguistic strategies used when disengaging from one's own moral agency. During Phase II, we will build a fully functional ITAAD and validate and test the tool in lab studies as well as on real-world social media data. Equally important to this effort is the continuation of our ongoing discussions around commercial applications within the domain of cyber warfare, misinformation and hate speech with stakeholders at Silicon Valley technology companies.