Mapping proven clinical threat assessment models (such as DASH, TRAP-18, and WAVR-21) to definable and detectable online behaviors. This work focuses on the practicality of deploying existing models and signals to online contexts, creating evidence backed baselines for online threat analysis, and highlighting necessary components for an actionable intervention.
Analyzing incidents of homicide with digital precursors to address bias toward content-based moderation. This study takes official reviews and recommendations made in the aftermath of violent events and evaluates the rate of suggested actions targetting content versus detection and intervention, as well as the potential for intervention based on manifest evidence of violent intent.
Evaluating existing data sources for behavioral threat analysis to highlight systemic gaps in availability and usage. This initiative includes proposals for alternative data strategies to improve internal platform detection and facilitate secure external data availability for safety research.
The SERAPH Project is always happy to collaborate on safety related projects. Preventing kinetic violence based on online signals inherently involves multiple stakeholders - online platforms identifying threats, law enforcement conducting escalations, regulators creating clear guidance and mandates, and NGOs and academics studying the most effective and impactful strategies for violence prevention.
In addition, one of our primary goals is to build and spread high quality resources and standards to improve safety across online spaces - especially smaller platforms that don't have the necessary data volume or the large dedicated teams needed for best-in-class detection and handling of incidents.
If you would like to discuss any of our research and initiatives, contribute to an existing project, or explore a collaboration in a related area of online threat analysis and kinetic violence prevention, please reach out via our contact page.