According to a new report from Media Matters, viral racism and anti-Semitic Tiktok videos appear to be made using VEO 3, Google’s new AI video generator.
The nonprofit research group found that some hate videos have received hundreds of thousands or millions of views.
A Tiktok spokesman told Mashable that the platform has corporate policies for hate speech and that it uses a comprehensive technology and conditioning process to implement them.
“We proactively enforce strong rules against speech and behavior and delete accounts we identified in the report, many of which have been banned before the report was published,” a spokesperson told Mashable in an email.
A tiktok marked “Atlanta’s Ordinary Waffle House” in which a restaurant with monkeys set up overspending, throwing watermelons and carrying fried chicken. When media affairs took a video screenshot, more than 622,000 views have been viewed.
Google’s VEO 3 AI video generator is different from anything you’ve ever seen. The world is not ready yet.
Some commentators confirmed the video’s racist stereotype. One person said, “All their behavior…to T…”
Another tiktok was uploaded in mid-June and had at least 835,000 views, prompted: “I asked the AI: ‘Ordinary Spiritual Aviation Experience.” The video also features monkeys, climbing on the plane.
Media Affairs said it determined that the video ran for up to eight seconds, the length of the video clip of Veo 3’s public text. These videos are also marked as “veo” in the corner, or use a hashtag, title or username related to VEO 3 or AI. They also include common mistakes, twists and nonsense texts for AI-generated videos.
Media Affairs released a compilation of clips it identified to its own YouTube account.
Mixable light speed
Mashable contacted Google for comments about media transaction reports, but received no response at the time of publication.
While the media focuses on videos appearing on Tiktok, some of the same offensive clips have also been posted to YouTube and Instagram.
However, the examples that can be viewed are much less involved, sometimes receiving only a few hundred likes and a few comments, suggesting that they are not as widely viewed as they were all the rage on Tiktok on Tiktok.
Generally speaking, videos created by weo 3 have hateful or racist content and become popular on other social media platforms, including Instagram.
When VEO 3 was released in late May, Mashable tech editor Timothy Beck Werth described its realism as “impressive” and “terrifying.” Google told WERTH that VEO 3’s error prevention information includes digital watermarks and adopts AI security guidelines.
Media Affairs identified AI-generated videos including anti-black stereotypes about crime, food preferences and absent fathers. Some people met with black people, including a white police officer shooting “black people” out of his car. The clip has been viewed over 14 million times.
The clips also portray racist images targeting the people of Asia and South Asia and depict anti-Semitic stereotypes, including Jews chasing gold coins.
A clip was watched a million times, and a gaunt man standing in front of the crematorium performed Vloggog at the same time as a Nazi concentration camp. The man said, “Well, everyone has had a great time here.” It’s not clear that the long video was made with VEO 3.
Another style of AI-generated video seems to focus on defending immigrants and protesters.
The videos appear to violate Google’s hate speech policy. Google’s Generative AI strategy prohibits users from generating or distributing content that promotes hateful or hate speech; harassing and abusing others, violent or inciting violence.
Tiktok prohibits hate speech and acts of hate “including attacks, threats, dehumanization, or dehumanizing individuals or groups based on their protected attributes.”
Updated: July 3, 2025, 1:02 PM PDT The update to the story includes Tiktok’s statement.
