Google’s Veo 3 video creator was released in May and represented a startling leap in AI video. The pixel-perfect output of the model can be used for nefarious reasons. While many viral AI videos are harmless fun, it is also possible to use the model for nefarious ends. Users have noticed an abundance of racist AI videos on TikTok. This app may or may not get banned in the next few months.
A report from In recent weeks, MediaMattersand numerous TikTok account have begun posting AI-generated videos using racist and antisemitic stereotypes. The AI vitriol is mostly directed at Black people. They are portrayed as “the usual suspects” criminals, parents who are absent, and monkeys who love watermelon. The content is also aimed at immigrants and Jews. The videos are eight seconds long and have the “Veo” logo, which confirms that they were created by Google’s AI model.
This compilation video contains examples from TikTok after the release of Veo 3. Be warned though, it contains racists and antisemitics content. The videos are shocking. This is probably the point. Nothing drives engagement on social networks like anger and drama. MediaMatters reported that the original posts had many comments echoing stereotypes used in video.
Veo 3 is spreading hateful AI videos on TikTok.
Google stressed security when it announced new AI models. We’ve all seen AIs refuse to complete tasks that violate their guardrails. It’s not fun when the system blocks your output despite your good intentions. Google has generally struck the right balance in the past, but Veo 3 appears to be more compliant. We tested a few simple instructions with Veo 3 to see if we could reproduce the videos.
