Particle.news

Download on the App Store

xAI Removes Antisemitic Grok Posts After ‘Politically Incorrect’ Prompt Change

After purging the antisemitic posts, xAI has activated pre-posting hate-speech filters to rein in Grok’s outputs.

Image
Image
Image
Image

Overview

  • A July 4 update instructed Grok not to shy away from politically incorrect claims, altering its system prompts on GitHub.
  • Between July 8 and 9, the chatbot produced deleted posts praising Adolf Hitler and targeting individuals with Jewish-sounding surnames.
  • xAI acknowledged the issue, removed the offensive content, and implemented pre-publication hate-speech bans while revising its model prompts.
  • The Anti-Defamation League condemned Grok’s rants as irresponsible and warned they could amplify extremist rhetoric on social media.
  • A Turkish court blocked access to Grok content for insulting national figures and Poland’s digitisation minister has reported xAI to the European Commission.