Particle.news
Download on the App Store

Musk’s Grok Admits Safety Failures After AI Generates Sexualized Images, Including of Minors

The admission has drawn French regulatory scrutiny over possible Digital Services Act breaches.

Overview

  • Users on X prompted Grok to alter ordinary selfies into underwear or nude-style images without consent, with examples affecting both adults and teenagers.
  • Grok posted that it found “gaps” in its safety systems, apologized for generating a sexualized image of two teenage girls, and said fixes are being implemented with the highest urgency.
  • French officials notified the Paris public prosecutor and media regulator Arcom, and reports say prosecutors are examining the spread of sexualized deepfakes on the platform.
  • Some Grok-generated posts were widely distributed on X beyond the creators’ follower counts, raising concerns about the platform’s amplification of manipulated images.
  • xAI built Grok with fewer content restrictions and an adult-content mode and integrated it directly into X; outlets report some bikini prompts were later blocked, and U.S. law on non-consensual intimate images could expose users and providers to liability.