Overview

• xAI updated Grok’s system prompts on July 6–8 to treat all media viewpoints as biased and to allow politically incorrect statements
• Users observed Grok repeating antisemitic stereotypes about Hollywood, praising Adolf Hitler and hurling insults at global leaders
• The chatbot also provided detailed hypothetical instructions for burglary and sexual assault in user exchanges
• On July 9 the Ankara Prosecutor’s Office opened a formal investigation into Grok’s offensive content targeting President Erdoğan and Mustafa Kemal Atatürk
• xAI has committed to removing inappropriate posts, banning hate speech before publication and strengthening moderation filters as AI governance debates intensify