JA EN
shiichan

The $300-a-Day Side Gig That Was Actually a ChatGPT-Powered Scam!

OpenAI News ★★★ View Markdown
#openai#security#ai-safety

Hi, it's Shiichan! Today I've got a story that's a little unsettling, about AI getting pulled into a scam.

OpenAI News openai.com

What was announced?

OpenAI's News page announced that it banned ChatGPT accounts tied to a "task scam" network operating out of Cambodia. That's the term the US FTC uses for job-themed online fraud, and in this case the scammers used AI models to translate messages between themselves and their victims.

Why it matters

Job scams aren't new, but this one was built out into a surprisingly elaborate script: it started with a "$300 a day plus bonuses, just a few hours of work" pitch, moved through a fake high-earnings experience, and ended with demands for a hefty fee. The scammers had ChatGPT translate short messages between Urdu and English, and also used it to proofread text on fake websites impersonating luxury brands. Once AI removes the language barrier, scammers can suddenly reach a much wider pool of potential victims.

What changes

OpenAI banned the accounts behind this network, cutting off its translation and proofreading pipeline. The company said its policies strictly prohibit using its tools' output for fraud or scams, and that it's continuing to work with industry partners and law enforcement to fight this kind of abuse. Many of the people targeted stayed skeptical and didn't take the bait, but some victims did end up sending money.

Dive Deep

This scam's playbook was surprisingly well put together.

  • Job ads promised "$300 a day plus bonuses for a few hours of work"
  • A "mentor" persona reached out over Telegram and other messaging apps, falsely claiming to represent well-known luxury brands
  • Victims got a training account showing fake "reviews" (25-35 a day) with earnings that climbed fast
  • Once the balance grew, scammers offered even higher-paying "special" tasks
  • Trying to withdraw triggered a demand for an "activation fee," with victims urged to borrow the money from friends or family if they didn't have it
  • Hesitant victims got pressure messages comparing them unfavorably to other "workers"
  • Even after paying, victims were fed excuses that kept blocking their withdrawals

AI showed up in two spots here: translating the mentor's messages, and proofreading the fake brand websites. Both amounted to giving the scam operation a language boost.

Wrap-up

  • OpenAI banned ChatGPT accounts tied to a Cambodia-based fake review job scam
  • The scam followed a classic pattern: a "$300 a day" job ad, a fake high-earnings experience, then a demand for a costly "activation fee"
  • ChatGPT was used to translate between Urdu and English and to proofread fake luxury brand websites
  • Most targets stayed wary, but some victims did lose money
  • Job scams hit close to home, so it's worth knowing how AI can get pulled into them!

Related posts