The entertainment world was shaken on a recent Tuesday night when Perez Hilton, the renowned blogger and media personality, livestreamed a distressing scene on TikTok. For approximately 15 minutes, Hilton appeared with a bloody torso and a knife, seemingly harming himself. This incident has reignited conversations about the efficacy of social media platforms in moderating live content.
The livestream was flagged by TikTok’s AI moderation system within minutes and sent to a human moderator for review. However, technical difficulties prevented the immediate removal of the stream. Despite these challenges, Hilton persisted, broadcasting on multiple streams. Miami law enforcement was alerted and arrived to find Hilton alone at his residence. He was subsequently transferred to a local hospital for medical attention.
Perez Hilton’s Livestream and the Challenges of Real-Time Moderation
The incident underscores the persistent difficulties social media companies face in moderating live content. Livestreaming became mainstream in 2015 when Facebook introduced its livestreaming feature. Initially, platforms relied heavily on user reports and human moderators to identify and remove inappropriate content. This approach proved to be mentally taxing for moderators, leading to numerous lawsuits from former employees who suffered trauma from prolonged exposure to graphic material.
Despite these early efforts, many violent livestreams managed to evade detection. Notable incidents include broadcasts of mass shootings and suicides. One particularly harrowing example was a 2016 livestream of a 12-year-old taking her own life, which remained on Facebook for nearly two weeks before being removed. The most significant wake-up call came in 2019 when a shooter in Christchurch, New Zealand, livestreamed an attack on two mosques, resulting in the deaths of 51 people.
The Evolution of Moderation Practices
In response to these incidents, social media platforms have adopted a hybrid approach combining AI and human moderation. AI now serves as the first line of defense, automatically flagging potentially problematic content. For instance, TikTok reported suspending around 50 million livestreams in the first quarter of 2026. However, the sheer volume of content—over 100 million videos uploaded daily on TikTok alone—poses a significant challenge.
Removing initial content is just one part of the equation. Moderators must also combat additional streams and subsequent recordings of the incident, a task akin to playing whack-a-mole. The situation is further complicated by layoffs in trust and safety teams, leaving platforms understaffed and struggling to keep up with the deluge of content.
The Aftermath of Perez Hilton’s Livestream
Following the incident, Hilton’s TikTok account was banned. However, as is often the case with social media, content from one platform quickly spreads to others. Unedited clips of Hilton’s livestream were widely shared on X (formerly Twitter), where they remained accessible for much of Wednesday morning. An X spokesperson did not respond to requests for comment.
Hilton has yet to publicly address the incident, but social media has been flooded with messages of support and concern from his fans. Alongside these well-wishes, there have been warnings urging users not to watch the graphic footage. The incident serves as a stark reminder of the ongoing challenges in ensuring the safety and well-being of users on social media platforms.
If you or someone you know is in crisis, please reach out to the Suicide & Crisis Lifeline at 988.
