New research tests four major platforms across 800 responses, exposing stark disparities between English and Persian language safety guardrails. | Read More in News from Israel
AI chatbots consistently fail to identify, reject antisemitism in Persian




