TikTok is facing renewed child-safety criticism after an investigation found that a 14-year-old account using the platform’s Restricted Mode could still be recommended graphic AI-generated cartoons depicting violence, sexual assault themes and torture. The investigation, published by The Mail on Sunday on August 30, 2026, found disturbing animated videos circulating through accounts connected to creators using tools from the startup Director AI. TikTok says Restricted Mode is intended to limit sexually suggestive content, realistic violence, weapons, drugs and other mature themes. After the company was contacted about the investigation, TikTok removed the specific videos cited by the newspaper but left the accounts that posted them active. [Mail on Sunday] [TikTok]
Investigators Create A 14-Year-Old TikTok Account
The Mail on Sunday tested TikTok using an account registered as belonging to a 14-year-old. Investigators activated Restricted Mode and spent hours watching material related to children’s television programs. [Mail on Sunday]
Restricted Mode was configured to limit categories including profanity, weapons, sexually suggestive material, drugs, violent or threatening imagery and mature themes. Those categories closely match TikTok’s own description of material that should generally be excluded from Restricted Mode. [TikTok]
Despite those settings, the account was recommended substantial amounts of content falling into those categories. [Mail on Sunday]
Graphic AI Cartoons Appeared In Restricted Mode
The investigation found animated videos depicting extreme violence. One cited video showed a cartoon doctor using a drill on a patient’s face during a grotesque cosmetic procedure intended to make the patient resemble Kim Kardashian. [Mail on Sunday]
Another video depicted an implied sexual assault following a burglary. The animation showed a tattooed burglar attacking a wealthy woman, leaving her battered and crying while he walked away with his pants unzipped. [Mail on Sunday]
Other content reviewed by investigators depicted torture and violent abuse. [Mail on Sunday]
The Animation Distinction Became Central To The Controversy
TikTok’s Restricted Mode documentation specifically identifies “realistic violence or threatening imagery” as content users should not expect to see. [TikTok]
The videos highlighted in the investigation were cartoons rather than realistic live-action footage. The investigation reported that animated depictions were able to circulate despite containing themes that would otherwise fall within categories Restricted Mode is intended to limit. [Mail on Sunday]
TikTok separately prohibits sexually explicit material in its own augmented-reality effect guidelines even when the material is animated. Its Effect House rules explicitly prohibit depictions of real or animated sexual activity, stimulation or arousal, as well as content glorifying sexual assault. [TikTok]
Director AI Is Connected To The Videos
The investigation focused on creators using technology developed by Director AI. The startup provides AI video-generation tools that allow creators to rapidly produce short animated and photorealistic videos. [Mail on Sunday]
Director AI is led by Ben Chiang, a former Snapchat director. Chiang has publicly described the company as developing an AI-powered video-creation agent. [LinkedIn]
Chiang said approximately 50 test users had generated tens of thousands of videos using Director AI tools. He wrote in 2026 that those videos had accumulated approximately 10 billion views across social platforms. [LinkedIn]
One AI Video Passed 100 Million Views
Chiang publicly highlighted one Director AI-generated video that received more than 100 million views in less than one week. [LinkedIn]
He described the phenomenon as “AI slop” while arguing that the dismissive label did not reflect how strongly audiences were engaging with the videos. [LinkedIn]
Chiang said the viral video had a completion rate above 50 percent. [LinkedIn]
Director AI Says Creators Control What They Publish
Director AI told the Mail that people using its software make their own editorial decisions. The company said creators decide what they produce and where they publish it, including whether they upload it to TikTok. [Mail on Sunday]
The company said it does not own the TikTok accounts distributing the videos. It also said it does not specifically pay those users to promote its companion app. [Mail on Sunday]
Chiang said some material created with Director AI tools may contain fictional mature themes. He said TikTok and other social networks are responsible for deciding what content qualifies for Restricted Mode under their own policies. [Mail on Sunday]
Reel AI Was Rated For Children As Young As Seven
Director AI’s associated Reel AI app was listed in Google Play with an age rating allowing children as young as seven to use it when the investigation was conducted. [Mail on Sunday]
The app allows users to continue viewing AI-generated stories and videos outside TikTok. Some TikTok creators promoted Reel AI as a destination for viewers wanting to see additional episodes or content. [Mail on Sunday]
Director AI changed the app’s age rating after being contacted by the newspaper. [Mail on Sunday]
TikTok Removes The Videos Cited In The Investigation
TikTok said its Community Guidelines prohibit content promoting or glorifying violence, sexual abuse and sexual exploitation. [Mail on Sunday]
After the Mail supplied TikTok with examples, the platform removed the videos specifically identified in the investigation. [Mail on Sunday]
The accounts responsible for posting them remained active when the newspaper published its report. Investigators said those accounts continued carrying other material they regarded as inappropriate for children. [Mail on Sunday]
TikTok Says Restricted Mode Uses Automation And Human Moderators
TikTok says Restricted Mode relies on both automated systems and human moderation. The company says its systems assess content and account behavior to determine what should be excluded. [TikTok]
TikTok says content can be excluded from Restricted Mode even if it does not violate the broader Community Guidelines. Material may remain on TikTok while still being withheld from users who have activated Restricted Mode. [TikTok]
The company asks users to report material they believe should not appear while Restricted Mode is enabled. [TikTok]
TikTok Previously Faced A Restricted Mode Pornography Investigation
The 2026 cartoon controversy follows an earlier investigation into sexual content reaching teen accounts. In 2025, Global Witness created seven UK TikTok accounts posing as 13-year-olds on factory-reset phones and enabled Restricted Mode. [Global Witness]
Researchers found TikTok’s search system recommending highly sexualized search terms. Those suggestions led the accounts toward pornographic or sexually explicit material within only a small number of interactions. [Global Witness] [Guardian]
The tests were conducted with no previous search history on the phones. Global Witness said the sexually explicit suggestions were produced by TikTok’s recommendation system rather than reflecting prior account behavior. [Global Witness]
TikTok Said It Took Action After The Earlier Investigation
Global Witness initially reported the search-suggestion issue to TikTok before publishing its investigation. According to the organization, TikTok responded that it had taken action globally. [Global Witness]
Researchers later repeated their tests and said the problem remained. The second round occurred after key child-protection requirements of Britain’s Online Safety Act had taken effect. [Global Witness]
U.S. States Have Previously Challenged TikTok’s Restricted Mode Claims
Indiana sued TikTok in 2022 over representations the company made about its safety features. The state alleged that Restricted Mode did not meaningfully prevent teenagers from encountering sexual, profane and otherwise mature material. [Indiana Complaint]
The Indiana complaint included tests involving 13-year-old accounts using Restricted Mode. Investigators alleged that sexual and profane content remained accessible and was sometimes recommended through the For You feed. [Indiana Complaint]
Iowa later brought similar allegations against TikTok. Its attorney general accused the platform of misleading parents about how effectively Restricted Mode filters mature material. [TechCrunch]
Pennsylvania Sues TikTok Over Child Safety
Pennsylvania Attorney General Dave Sunday filed a new lawsuit against TikTok in August 2026. The state alleges TikTok misrepresents the amount of mature content children can encounter while maintaining an app-store age rating suitable for teenagers. [Pennsylvania AG]
The Pennsylvania complaint specifically addresses sexual content, nudity, profanity and drug-related material. [Pennsylvania AG]
Sunday has also criticized Restricted Mode and Family Pairing. His office alleges that the tools can fail to shield young users from material parents reasonably expect them to block. [WVIA]
Ofcom Is Investigating TikTok’s Child-Safety Systems
Britain’s communications regulator opened a formal TikTok investigation on July 16, 2026. Ofcom is examining whether the company is complying with its legal obligations to protect children from harmful content under the Online Safety Act. [Ofcom]
The law requires platforms likely to be accessed by children to use proportionate systems designed to prevent minors from encountering specified categories of harmful material. [Ofcom]
The UK government said after the AI-cartoon investigation that Ofcom had its full backing to enforce those requirements. [Mail on Sunday]
European Commission Also Raises TikTok Child-Safety Concerns
The European Commission issued preliminary findings against TikTok in July 2026 under the Digital Services Act. The Commission said TikTok had not provided sufficiently safe default account settings for minors. [European Commission]
The Commission separately found preliminarily in February that TikTok’s addictive design violated the Digital Services Act. Regulators cited features including infinite scroll, autoplay, push notifications and the personalized recommendation system. [European Commission]
TikTok Reached A $400 Million U.S. Children’s Privacy Settlement
TikTok and ByteDance agreed in August 2026 to pay $400 million to resolve U.S. government litigation involving children’s privacy. [DOJ]
The lawsuit concerned compliance with the Children’s Online Privacy Protection Act. The Justice Department said TikTok had since implemented extensive changes involving age controls, parental oversight and privacy protections. [DOJ]
The settlement did not constitute a finding that TikTok was liable for the government’s allegations. [DOJ]
One of the first political bloggers in the world, Oliver Willis has operated OliverWillis.com since 2000. Contributor at Media Matters for America and The American Independent. Follow on Twitter at @owillis. Full bio.