S'porean Man Used ChatGPT to Plan Honey Trap, Extort S$20,000
Source: Mothership
A 20-year-old Singaporean man turned to ChatGPT to plan a honey trap scheme, posing as a teenage girl online to extort S$20,000 from another young man. Anthony Xu (transliterated from Mandarin) pleaded guilty to one count of extortion and one count of cheating, and is now awaiting sentencing. The sc
A 20-year-old Singaporean man turned to ChatGPT to plan a honey trap scheme, posing as a teenage girl online to extort S$20,000 from another young man. Anthony Xu (transliterated from Mandarin) pleaded guilty to one count of extortion and one count of cheating, and is now awaiting sentencing.
The scheme began in October last year, when Xu learned from a female friend that a 21-year-old man had been persistently trying to chat her up on social media. Drawing on Instagram short videos and ChatGPT for ideas, Xu built a fake persona — a 14-year-old girl — and flirted with the victim online before arranging to meet at an MRT station. After the meet-up, Xu claimed to be the "girl's" older brother, threatened the victim, filmed him, photographed his identity card, and warned he would alert station staff and the police. The intimidation continued in Toa Payoh until the victim broke down and agreed to pay S$20,000 in "hush money" at S$300 a month.
In the end, only S$100 changed hands. The court heard that Xu felt guilty after reflecting on what he had done, returned the money the next day, and apologised to the victim's mother — unaware that a police report had already been filed. He was arrested the following day. His lawyer said in mitigation that Xu cooperated fully with investigations, pleaded guilty early, and showed genuine remorse, asking the court for leniency. The judge has ordered a report on Xu's suitability for probation and reformative training, with sentencing set for Sep 23.
Why it matters for Singapore: This case shows that generative AI has moved into the scammer's toolkit, not just the defender's. Singapore runs some of the world's most aggressive anti-scam campaigns — ScamShield, police advisories, and bank-level safeguards — but AI makes it cheaper than ever to spin up convincing fake personas at scale. The practical takeaway for users is blunt: verify who you are actually talking to online, especially when money or intimacy enters the picture. For regulators and educators, it is another sign that scam awareness and AI literacy need to keep evolving as fast as the tools do.
