Various accounts are using now AI TikTok video generator tools to create generated content that includes overtly racist and anti-immigrant material, attracting billions of views, according to Paris-based non-profit, AI Forensics.
The newly risen TikTok synthetic media, including hyper-realistic face-altering filters, is now weaponized online to erode social trust and spread misinformation like wildfire, on social platforms originally algorithmically optimized for virality.
According to the report, AI generated TikTok videos are now spreading at a speed that no human operation could ever achieve, flooding feeds with synthetic faces, political messages, and manipulated imagery.
With this, comes a deeper risk exposing the disadvantages of AI in social media. Where generative AI influences common beliefs, long before realizing the artificiality of the content.
TikTok Synthetic Media Policy AI-Generated Content Labeling 2025
In just one month, researchers identified 354 automated accounts posting 43,000 videos. many created with AI TikTok clip maker that allows users to mass-produce content.
Some accounts posted up to 70 times a day, often at the same time, showing automated behavior that exploits TikTok’s system recommendations.
Much of the content featured sexualized synthetic women, adding to a broader surge of AI TikTok videos mimicking real people with alarming realism. According to the report, AI women are always stereotypically attractive, with sexualized cleavage, with some holding a resemblance to girls who appeared underage.
TikTok describes the findings as “unsubstantiated,” but many accounts escaped detection for months. Their removal came only after researchers and journalists flagged them.
Other accounts pushed political misinformation through AI generated TikTok videos, including computer generated news segments modeled after major broadcasters. These, combined with fabricated visuals, are examples of how generative content weaponizes bias at an algorithmic scale.
GAN-Powered Fakes and Public Perception Online
The fast rise of generative AI social media tools is tied to the same technology behind advanced Generative Adversarial Networks (GAN) systems, which create faces and voices that never existed. GANs work by letting two AI models compete until the fake becomes almost impossible to distinguish from reality.
This technique is already used in filters that reshape faces, boost beauty features, or alter age, often without users noticing. The risk goes far beyond cosmetic editing; highly realistic synthetic identities can be mixed with biased narratives, turning manipulated visuals into persuasive messages.
Synthetic Influence and Trust Destruction
AI Forensics also exposed videos associated with TikTok AI memes about immigration that use comic forms for the expression of injurious stereotypes. In many, a synthesized voice of a man using hate speech appeared on TikTok in artificial narrations that took on authority.
Beyond harmful themes, researchers noted a rise in “slop” content, where bizarre AI tools to create TikTok videos meant only to grab attention.
Even these can influence perception when repeated endlessly, especially when produced using the AI TikTok video maker, or when reshared into broader ecosystems.
Some of the most disturbing content included deepfake sexual content, illustrating the proliferation of TikTok deepfake porn in violation of the platform’s rules but still spreading through gaps in detection. These failures reflect broader weaknesses of AI on social media, where moderation systems cannot keep pace with synthetic media.
The study warns that platforms are now a part of the global generative AI social media landscape in which synthetic narratives look more real than reality.
“The blurring line between authentic human and synthetic AI-generated content on the platform is signaling a new turn towards more AI-generated content on users’ feeds,” according to the researchers.
The investigation by The Guardian highlighted AI’s impact on social media algorithms as increasingly prioritizing engagement over accuracy, birthing TikTok algorithmic radicalization, were repeated synthetic messages shape beliefs.
AI Forensics welcomed TikTok’s new AI-reduction tools but remains skeptical, mentioning that labelling must become visible and consistent if it is to protect users. Without stronger safeguards, the AI TikTok Video Generator and similar systems may continue to reshape public opinion without users realizing it.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Tech sections to stay informed and up-to-date with our daily articles.