Cambodian scammers use AI to deceive victims across Asia-Pacific region
Meta reveals Cambodian scammers using AI, deepfakes, and tailored scams across Asia-Pacific, removing 6,400+ fraudulent items to protect users
- Meta reveals Cambodian scam networks using AI to deceive victims across Asia-Pacific
- Scammers often impersonate government officials and police
- AI technology is used to enhance scams, such as creating deepfake videos with famous individuals’ images, generating multilingual job recruitment scams, and customizing communication with victims
- These scam networks have become increasingly complex and widespread, using stolen or fabricated personal data, encrypted chat apps, and automated systems to expand operations
Meta has released its bi-annual Adversarial Threat Report, summarizing major global changes regarding online scams, influence operations, and AI-driven violations. One key finding is the handling of Cambodian criminal networks impersonating government officials and police to deceive victims across the Asia-Pacific region.
Over 6,400 items (accounts or other content) were removed between January and October 2025 in countries including Australia, Indonesia, the Philippines, Thailand, and Vietnam.
Moreover, Meta observed an increase in the size and complexity of cross-border scam networks in the region, including the following key details:
The scam networks in the Asia-Pacific have developed in scale and complexity: With scam centers pushing victims to use encrypted chat apps and alternative payment channels. Stolen or fabricated identities are used to make the scams appear more credible in specific countries, and operations are expanded through templates and automation. These strategies have been consistently observed in Australia, Singapore, and India, where impersonation and investment scams remain the most reported threats.
AI enhances both scam and prevention efforts: Scammers in the Asia-Pacific region are using AI to create deepfake videos with famous individuals’ images (a common practice in Australia and India), generate multilingual job recruitment scams, and tailor communication with each victim to increase effectiveness.
In response, Meta has employed advanced behavior detection systems, suspicious interaction alerts, internal threat controls, and new AI safeguards like Llama Firewall and the Rule of Two framework for AI agents.
Meta’s global and regional efforts to protect people have also supported the Singapore Police’s arrests related to scam networks, in collaboration with law enforcement across the region.
Nearly 500,000 public figures have participated in Meta’s facial recognition program to combat impersonation scams, which have been prevalent in Australia, Singapore, and India.
In the first half of 2025, reports of fraudulent ads using celebrities as bait decreased by 22% globally compared to total ad impressions.
Meta continues to delete large numbers of fake accounts and collaborate with law enforcement agencies, financial institutions, and cross-sector partners through programs such as the U.S. Department of Justice’s (DOJ) Scam Center Strike Force.
In Thailand, Meta has actively fought against online fraud in the first half of 2025 by addressing content, accounts, and ads violating fraud and deception policies.
Key actions include:
- Removing over 7.3 million pieces of content on Facebook and over 15,000 on Instagram, with 99% of Facebook content and 76% of Instagram content removed before being reported
- Deleting over 66,000 Facebook accounts and more than 24,000 Instagram accounts, with 65% of Facebook accounts and 86% of Instagram accounts deleted before being reported
- Banning over 18,000 Marketplace listings on Facebook, with 84% removed before being reported
- Removing over 380,000 fraudulent ads, with more than 90% deleted before being reported
- Banning over 3,000 ad accounts before being reported
