Most people would log off after days of online harassment. Singapore content creator Renae Cheng endured it for more than two years before speaking up — and what she described sounds like a nightmare scripted by artificial intelligence.
In a video that ran nearly three minutes, Cheng detailed a campaign of harassment that began in 2024: multiple burner accounts across Instagram, TikTok and even Gmail, long hateful messages, threats of harm. The harasser, she said, is a woman based in the UK. The worst of it was AI-generated videos depicting Cheng in compromising situations — including being groped by a man who is not her husband.
The harassment spilled beyond Cheng herself. Her husband was targeted, and her eight-month-old baby was called a "jin" — which she was told means "demon." When she reported the AI videos to TikTok, the harasser retaliated by spamming screenshots across her posts.
The origin story, Cheng said, is "borderline laughable." It allegedly began after a 2023 TikTok video in which she reviewed Eczema Mittens, a product for the skin condition she has lived with for years. The business owner is a Singaporean man named Reuben — the same man she was made to make out with and be groped by in the fabricated videos. According to Cheng, the harasser had first messaged her professing a crush on the business owner and asking for his contact details; she appears to suffer from eczema and topical steroid withdrawal herself, which Cheng said made the fixation even more puzzling. The woman, Cheng added, runs a dedicated TikTok account posting AI videos of the man "canoodling" with anyone he collaborates with.
Cheng's decision to go public turned a private torment into a public conversation about AI-enabled abuse — deepfakes as a weapon not of geopolitics but of personal vendetta, aimed at an ordinary creator with a baby at home. Singapore's laws against harassment and doxxing are among the region's toughest; whether they can reach a harasser in the UK is the harder question.
Cheng's ordeal sits at the intersection of two accelerating trends: the weaponisation of generative AI and the inadequacy of existing harassment laws. AI-generated non-consensual imagery — once the preserve of celebrities — is now a tool anyone can deploy against anyone, with free tools capable of producing convincing fakes in minutes. Singapore's Protection from Harassment Act criminalises doxxing and threatening behaviour, and recent amendments have strengthened protections against online abuse, but enforcement across borders remains the hard part.
The cross-border dimension is what makes cases like Cheng's so intractable. Her alleged harasser is in the UK; content is hosted on global platforms; Singaporean police must navigate mutual legal assistance channels that move at bureaucratic speed. TikTok's content moderation — reactive, inconsistent, easily gamed by burner accounts — offered Cheng little protection over two years.
By going public, Cheng did what the system could not: she named the behaviour, showed its human cost, and forced a conversation about AI abuse that usually stays abstract. Her baby being called a "demon" by a stranger's burner account is the detail that lingers — a reminder that behind every discussion of "AI safety" are real people absorbing the damage.