AI-generated racist videos are spreading widely on TikTok

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

A nonprofit group called Media Matters has found a troubling trend: racist videos that look like they were made with Google’s AI video tool, Veo 3, are getting millions of views on TikTok. 

These AI-generated videos contain hateful stereotypes, especially targeting Black people, and are spreading fast on social media.

Media Matters figured out the videos were made with Veo 3 because they had a “Veo” watermark visible in the corner. 

Some users who posted these clips even included hashtags, captions, or usernames mentioning Veo 3 or AI, making it clear they used Google’s tool. 

The videos were very short, lasting eight seconds or less, which fits with Veo 3’s limit on video length.

Google launched Veo 3 in May, saying it would let people create videos and audio using just a text prompt. On Google’s website, the company promises to block harmful or abusive requests and results. 

Similarly, TikTok’s own policies state that hate speech and hateful content are not allowed, and videos with negative stereotypes about people because of their race or other protected characteristics shouldn’t be recommended or promoted on the app.

A TikTok spokesperson responded to the report, saying the company takes action against hate speech and had already banned many of the accounts mentioned by Media Matters before the report even came out. 

They emphasized TikTok’s commitment to removing hateful content and enforcing its rules.

However, The Verge found that some of these racist videos were also on YouTube, though they had fewer views there. 

Meanwhile, Wired discovered similar offensive AI-generated videos circulating on Instagram, too. 

Besides anti-Black racism, Media Matters also found AI videos spreading antisemitic stereotypes and hateful depictions of immigrants and Asian people.

The findings show that despite promises from big tech companies like Google and TikTok to prevent hateful content, AI tools can still be misused to create and share racist videos, and those videos can quickly rack up millions of views online, reaching huge audiences before being taken down.

Are you even surprised at this point? What do you think social media companies need to do better to combat misinformation and hateful content? Tell us below in the comments, or via our Twitter or Facebook.