A safety feature on the social media platform TikTok, designed to shield children from inappropriate content, has failed to prevent the exposure of graphic and disturbing videos to underage users, according to an investigation. The app’s “restricted mode,” which is intended to filter out harmful material, has been circumvented by users exploiting loopholes in TikTok’s policies, allowing explicit animated content depicting violence and sexual abuse to reach young audiences.

The controversy centers on AI-generated videos produced using tools developed by a startup called Director AI, founded last year by former Snapchat director Ben Chiang and two colleagues. These tools have facilitated the creation of more than 15,000 violent and distressing clips, some of which have garnered significant viewership on TikTok. One widely viewed video features an animated doctor performing a gruesome operation, while another depicts a tattooed burglar attacking a wealthy woman and rifling through her belongings. The content is described as algorithmically engineered to be addictive.

Director AI offers its videos through a companion app called Reel AI, which was until recently rated suitable for children as young as seven. This app allows users to watch a limited number of free episodes of each mini-series before charging for additional content. One paid episode includes a scene where a mother uses a hairdryer to electrocute her baby. Following inquiries, the app’s age rating has since been revised.

In response to the findings, Shadow Education Secretary Laura Trott criticized the exposure of children to such material online, emphasizing that protections must be in place comparable to those offline. She urged lawmakers, particularly from the Labour Party, to take stronger action to safeguard children on social media platforms.

Ben Chiang acknowledged that Director AI provides the video-creation tools, but asserted that the content posted on TikTok is controlled by individual creators who decide what to publish and are not directly paid by Director AI to promote the app. He admitted that some of the videos contain mature themes but pointed to TikTok’s responsibility in applying its own content filtering policies.

TikTok reiterated that its community guidelines prohibit content that promotes or glorifies violence and sexual abuse. The platform said it has removed the videos cited in the report but noted that the accounts responsible remain active, continuing to share other questionable material.

Meanwhile, the Department for Culture, Media and Sport emphasized that Ofcom, the UK’s broadcast regulator, has full government backing to enforce the Online Safety Act. The department underscored the legal requirements for tech companies to remove illegal content and protect children from harmful material.