On Tuesday night, Perez Hilton, real name Mario Lavandeira Jr., appeared on a TikTok livestream naked and covered in a red fluid resembling blood, holding what looked like a knife. The stream prompted multiple viewers to alert authorities, leading to a crisis response unit and mental health professionals being dispatched to his home by the Miami-Dade County Sheriff's Office. According to detective Samantha Choon, the office received several calls about someone livestreaming acts of self-harm. A public information officer confirmed to WIRED that the individual was Lavandeira.
TikTok spokesperson Jamie Favazza stated that automated systems flagged the stream minutes after it began, but a "moderator error" delayed its removal and the banning of the account. The company claims the livestream lasted approximately 15 minutes and that it contacted law enforcement soon after detection. Favazza denied claims the stream remained online for 30 minutes or more, saying subsequent broadcasts by Lavandeira were cut within 90 seconds. TikTok's content policies prohibit content promoting suicide, self-harm, or graphic violence, and the platform uses both automated tools and human moderators to enforce these rules.
The incident follows past controversies involving livestreamed self-harm on social platforms, including the 2020 case of a Brazilian man and the spread of Ronnie McNutt's suicide footage. TikTok entered a new US ownership structure in January, with a joint venture now overseeing content moderation and data security. Dante Rusciolelli, Lavandeira's manager and CEO of Golden Artists Entertainment, said the team could not comment on TikTok's moderation decisions or communications with law enforcement. He emphasized that their focus remains on Lavandeira's health and requested privacy during his recovery.
TikTok's claim of a "moderator error" raises questions about the reliability of its content review process, especially given the 15-minute delay despite automated detection. The platform's ability to enforce its own policies during real-time crises appears inconsistent, even as it asserts cooperation with law enforcement. Lavandeira's manager declined to address moderation failures, focusing instead on privacy and care.
Editorial note: AI-assisted opinion, not established fact. Full disclaimer →
