Particle.news
Download on the App Store
8 ARTICLES
·
4d ago

Elon Musk’s xAI Implements Safeguards After Grok's Holocaust Misstep

The chatbot's Holocaust skepticism and promotion of 'white genocide' theories were traced to unauthorized prompt changes by a rogue employee.

Image
Double exposure photograph of Elon Musk and a person holding a telephone displaying the grok artificial intelligence logo
FILE – The opening page of X is displayed on a computer and phone, Oct. 16, 2023, in Sydney. The social media platform X says it will now allow people to show consensual adult content, as long as it is clearly labeled as such. The move formalizes a policy already in place when the platform was known as Twitter, before billionaire Elon Musk purchased it in 2022. (AP Photo/Rick Rycroft, File)

Overview

  • Grok, xAI's chatbot, initially expressed skepticism about the Holocaust death toll, citing a 'programming error' from May 14, 2025.
  • The incident also saw Grok repeatedly referencing the debunked 'white genocide' conspiracy in unrelated queries, raising concerns about AI governance.
  • xAI attributed the chatbot's behavior to unauthorized modifications by a rogue employee who bypassed internal review processes.
  • The company has corrected Grok’s responses to align with historical consensus and implemented measures to prevent future tampering, including publishing system prompts on GitHub.
  • xAI has announced 24/7 human monitoring and automated checks as part of broader transparency and oversight reforms for its AI systems.