Racist videos made with AI seen by millions on TikTok and YouTube

Google’s AI-powered video-generating tool Veo 3 was used to create racist videos that have garnered millions of views on TikTok and YouTube, Media Matters reports. The watchdog says the videos feature the Veo watermark in the corner and that one has received 14.2 million views. Some users list hashtags, captions or even usernames that belong to Veo 3.

The video clips are either all eight seconds long or are made up of multiple clips that are just eight seconds long, which aligns with the Veo 3 video generator's limit, The Verge explains.

Google's solution, launched in May, states that it "blocks harmful requests and results" and TikTok also states in its rules that the platform "does not recommend content that contains negative stereotypes about a person or group with protected attributes". Therefore, it is not yet clear how the videos were, firstly, generated by Google's tool and, secondly, approved to be visible on the Chinese platform.

TikTok spokeswoman Ariane de Selliers said: “We proactively enforce robust rules against hate speech and behavior and have removed accounts identified in the report, many of which were already banned before the report was published.” Google has yet to officially respond to the report, but similar videos have been discovered on YouTube and Instagram.