Most harmful content is implicit or coded. Current detection systems miss the layer where most harm actually lives.
AddressHate is building the research rigor and precision tools to surface digital hate proactively, and equip those positioned to act.
AddressHate is purpose-built to be different from what exists today. The interdisciplinary field is large and well-funded, but every existing approach catches hate after it has already surfaced. AddressHate is built for what current tools cannot see — detecting, mapping, and disrupting hate before it surfaces.
Linguistics, computer science, history, sociology, and policy under one roof. The problem is interdisciplinary, and so are our methods.
Peer-reviewed methods, expert annotation, and transparent benchmarks. Our work is not black-box detection trained on opaque labels.
Research becomes tools. Tools reach practitioners. Practitioners change outcomes. Every project is measured against that standard.