1 paper · 1 filter
Evžen Wybitul
AI safety systems face the dual-use dilemma. It is unclear whether to answer dual-use requests, since the same query could be either harmless or harmful depending on who made it an…