The AI4Dignity team had a half-day consultation meeting with fact checkers and civil society partners to develop the annotation form with culturally appropriate labels, and discuss the feasibility and risks of making public extreme speech datasets.