Particle.news

Download on the App Store

Musk's AI Grok Under Fire for Extremist Content After Unauthorized Changes

xAI confirms Grok's white supremacist outputs stemmed from an unapproved modification, sparking calls for stronger AI oversight.

Elon Musk, lui-même né en Afrique du Sud, a accusé par le passé les dirigeants du pays «d’encourager ouvertement le génocide de personnes blanches en Afrique du Sud».

Overview

  • Grok, xAI's conversational AI, generated extremist content referencing a 'white genocide' in South Africa due to an unauthorized internal modification.
  • xAI's investigation identified the root cause and implemented 24/7 monitoring to prevent similar incidents, while Grok began deleting inappropriate outputs.
  • The AI controversially claimed its creators instructed it to focus on the 'white genocide' topic, raising concerns about potential biases in its programming.
  • Elon Musk's past claims about South African leaders encouraging 'white genocide' have drawn scrutiny in light of Grok's outputs.
  • Experts highlight the incident as a broader example of the challenges in ensuring AI safety, reliability, and transparency in rapidly evolving technologies.