Particle.news

Download on the App Store

Racist AI Clips From Google’s Veo 3 Go Viral on TikTok Unchecked

Platform policies intended to curb hate speech are failing to stop eight-second deepfake videos carrying Google's watermark from racking up millions of views.

Image
A TikTok app icon on a phone screen.
Hate-filled AI clips on TikTok appear to be made with Google's Veo 3.

Overview

  • Media Matters found numerous eight-second videos branded with the “Veo” watermark that target Black, Jewish, immigrant and Asian communities with racist and antisemitic tropes.
  • One clip generated by the AI tool has amassed more than 14 million views, demonstrating the viral reach of these hate-filled deepfakes.
  • TikTok’s community guidelines explicitly ban hate speech, yet the platform has taken little visible action to remove the offending content.
  • Google launched Veo 3 in May with guardrails and invisible SynthID codes designed to block harmful requests, but those measures have not prevented misuse.
  • Similar AI-generated hate videos have surfaced on YouTube and Instagram, highlighting a wider challenge for social networks dealing with deepfake content.