Summary

  • Grok, the ChatGPT competitor owned by Elon Musk, stopped responding to certain queries after an ex-OpenAI employee at xAI updated its system prompt excluding results relating to Musk and former US President Donald Trump.
  • xAI’s engineering head, Igor Babuschkin, confirmed the incident, but claimed that it violated the company’s values and that the system prompt was meant to be transparent for users.
  • Babuschkin blamed an unnamed ex-OpenAI employee for the unauthorized change, which was supposedly made to prevent “spread of misinformation.”
  • Grok reportedly criticised Musk and Trump and stated that they were “doing the most harm to America,” among other things — potentially explaining the motivation for the censorship.

By Wes Davis

Original Article